maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[X86] Split more masked instruction tests to enable D60940. 

More mechanical splitting of tests so we can add a one use
check to the isel patterns for forming masked instructions.

In a few cases I changed immediates of instructions in
order to avoid needing to split.
This commit is contained in:
Craig Topper 2020-03-25 09:23:07 -07:00
parent d622612e61
commit 31c5afb3f2
25 changed files with 5010 additions and 4154 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

265
llvm/test/CodeGen/X86/avx512-gfni-intrinsics.ll
View File

@ -8,8 +8,8 @@ define <16 x i8> @test_vgf2p8affineinvqb_128(<16 x i8> %src1, <16 x i8> %src2, <
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xcf,0xd9,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x03]
; X86-NEXT: vgf2p8affineinvqb $4, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x04]
; X86-NEXT: vgf2p8affineinvqb $5, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x05]
; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -18,18 +18,20 @@ define <16 x i8> @test_vgf2p8affineinvqb_128(<16 x i8> %src1, <16 x i8> %src2, <
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xcf,0xd9,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x03]
; X64-NEXT: vgf2p8affineinvqb $4, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x04]
; X64-NEXT: vgf2p8affineinvqb $5, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x05]
; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8affineinvqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 3)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
%4 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> %passthru
%5 = xor <16 x i8> %3, %4
%6 = xor <16 x i8> %5, %2
ret <16 x i8> %6
%3 = call <16 x i8> @llvm.x86.vgf2p8affineinvqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 4)
%4 = call <16 x i8> @llvm.x86.vgf2p8affineinvqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 5)
%5 = select <16 x i1> %1, <16 x i8> %3, <16 x i8> zeroinitializer
%6 = select <16 x i1> %1, <16 x i8> %4, <16 x i8> %passthru
%7 = xor <16 x i8> %5, %6
%8 = xor <16 x i8> %7, %2
ret <16 x i8> %8
}
declare <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8>, <32 x i8>, i8)
@ -38,8 +40,8 @@ define <32 x i8> @test_vgf2p8affineinvqb_256(<32 x i8> %src1, <32 x i8> %src2, <
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xcf,0xd9,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x03]
; X86-NEXT: vgf2p8affineinvqb $4, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x04]
; X86-NEXT: vgf2p8affineinvqb $5, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x05]
; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -48,18 +50,20 @@ define <32 x i8> @test_vgf2p8affineinvqb_256(<32 x i8> %src1, <32 x i8> %src2, <
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xcf,0xd9,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x03]
; X64-NEXT: vgf2p8affineinvqb $4, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x04]
; X64-NEXT: vgf2p8affineinvqb $5, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x05]
; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 3)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
%4 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> %passthru
%5 = xor <32 x i8> %3, %4
%6 = xor <32 x i8> %5, %2
ret <32 x i8> %6
%3 = call <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 4)
%4 = call <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 5)
%5 = select <32 x i1> %1, <32 x i8> %3, <32 x i8> zeroinitializer
%6 = select <32 x i1> %1, <32 x i8> %4, <32 x i8> %passthru
%7 = xor <32 x i8> %5, %6
%8 = xor <32 x i8> %7, %2
ret <32 x i8> %8
}
declare <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8>, <64 x i8>, i8)
@ -68,8 +72,8 @@ define <64 x i8> @test_vgf2p8affineinvqb_512(<64 x i8> %src1, <64 x i8> %src2, <
; X86: # %bb.0:
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xcf,0xd9,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x03]
; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x03]
; X86-NEXT: vgf2p8affineinvqb $4, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x04]
; X86-NEXT: vgf2p8affineinvqb $5, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x05]
; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -78,18 +82,20 @@ define <64 x i8> @test_vgf2p8affineinvqb_512(<64 x i8> %src1, <64 x i8> %src2, <
; X64: # %bb.0:
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xcf,0xd9,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x03]
; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x03]
; X64-NEXT: vgf2p8affineinvqb $4, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x04]
; X64-NEXT: vgf2p8affineinvqb $5, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x05]
; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 3)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
%4 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> %passthru
%5 = xor <64 x i8> %3, %4
%6 = xor <64 x i8> %5, %2
ret <64 x i8> %6
%3 = call <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 4)
%4 = call <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 5)
%5 = select <64 x i1> %1, <64 x i8> %3, <64 x i8> zeroinitializer
%6 = select <64 x i1> %1, <64 x i8> %4, <64 x i8> %passthru
%7 = xor <64 x i8> %5, %6
%8 = xor <64 x i8> %7, %2
ret <64 x i8> %8
}
declare <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8>, <16 x i8>, i8)
@ -98,8 +104,8 @@ define <16 x i8> @test_vgf2p8affineqb_128(<16 x i8> %src1, <16 x i8> %src2, <16
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xce,0xd9,0x03]
; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x03]
; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x03]
; X86-NEXT: vgf2p8affineqb $4, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x04]
; X86-NEXT: vgf2p8affineqb $5, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x05]
; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -108,18 +114,20 @@ define <16 x i8> @test_vgf2p8affineqb_128(<16 x i8> %src1, <16 x i8> %src2, <16
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xce,0xd9,0x03]
; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x03]
; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x03]
; X64-NEXT: vgf2p8affineqb $4, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x04]
; X64-NEXT: vgf2p8affineqb $5, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x05]
; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 3)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
%4 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> %passthru
%5 = xor <16 x i8> %3, %4
%6 = xor <16 x i8> %5, %2
ret <16 x i8> %6
%3 = call <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 4)
%4 = call <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 5)
%5 = select <16 x i1> %1, <16 x i8> %3, <16 x i8> zeroinitializer
%6 = select <16 x i1> %1, <16 x i8> %4, <16 x i8> %passthru
%7 = xor <16 x i8> %5, %6
%8 = xor <16 x i8> %7, %2
ret <16 x i8> %8
}
declare <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8>, <32 x i8>, i8)
@ -128,8 +136,8 @@ define <32 x i8> @test_vgf2p8affineqb_256(<32 x i8> %src1, <32 x i8> %src2, <32
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xce,0xd9,0x03]
; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x03]
; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x03]
; X86-NEXT: vgf2p8affineqb $4, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x04]
; X86-NEXT: vgf2p8affineqb $5, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x05]
; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -138,18 +146,20 @@ define <32 x i8> @test_vgf2p8affineqb_256(<32 x i8> %src1, <32 x i8> %src2, <32
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xce,0xd9,0x03]
; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x03]
; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x03]
; X64-NEXT: vgf2p8affineqb $4, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x04]
; X64-NEXT: vgf2p8affineqb $5, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x05]
; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 3)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
%4 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> %passthru
%5 = xor <32 x i8> %3, %4
%6 = xor <32 x i8> %5, %2
ret <32 x i8> %6
%3 = call <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 4)
%4 = call <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 5)
%5 = select <32 x i1> %1, <32 x i8> %3, <32 x i8> zeroinitializer
%6 = select <32 x i1> %1, <32 x i8> %4, <32 x i8> %passthru
%7 = xor <32 x i8> %5, %6
%8 = xor <32 x i8> %7, %2
ret <32 x i8> %8
}
declare <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8>, <64 x i8>, i8)
@ -158,8 +168,8 @@ define <64 x i8> @test_vgf2p8affineqb_512(<64 x i8> %src1, <64 x i8> %src2, <64
; X86: # %bb.0:
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xce,0xd9,0x03]
; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x03]
; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x03]
; X86-NEXT: vgf2p8affineqb $4, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x04]
; X86-NEXT: vgf2p8affineqb $5, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x05]
; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -168,107 +178,162 @@ define <64 x i8> @test_vgf2p8affineqb_512(<64 x i8> %src1, <64 x i8> %src2, <64
; X64: # %bb.0:
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xce,0xd9,0x03]
; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x03]
; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x03]
; X64-NEXT: vgf2p8affineqb $4, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x04]
; X64-NEXT: vgf2p8affineqb $5, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x05]
; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 3)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
%4 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> %passthru
%5 = xor <64 x i8> %3, %4
%6 = xor <64 x i8> %5, %2
ret <64 x i8> %6
%3 = call <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 4)
%4 = call <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 5)
%5 = select <64 x i1> %1, <64 x i8> %3, <64 x i8> zeroinitializer
%6 = select <64 x i1> %1, <64 x i8> %4, <64 x i8> %passthru
%7 = xor <64 x i8> %5, %6
%8 = xor <64 x i8> %7, %2
ret <64 x i8> %8
}
declare <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8>, <16 x i8>)
define <16 x i8> @test_vgf2p8mulb_128(<16 x i8> %src1, <16 x i8> %src2, <16 x i8> %passthru, i16 %mask) {
; X86-LABEL: test_vgf2p8mulb_128:
define <16 x i8> @test_vgf2p8mulb_128(<16 x i8> %src1, <16 x i8> %src2) {
; CHECK-LABEL: test_vgf2p8mulb_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8> %src1, <16 x i8> %src2)
ret <16 x i8> %1
}
define <16 x i8> @test_vgf2p8mulb_128_mask(<16 x i8> %src1, <16 x i8> %src2, <16 x i8> %passthru, i16 %mask) {
; X86-LABEL: test_vgf2p8mulb_128_mask:
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xd9]
; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xe1]
; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xcf,0xd1]
; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_128:
; X64-LABEL: test_vgf2p8mulb_128_mask:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xd9]
; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xe1]
; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xcf,0xd1]
; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8> %src1, <16 x i8> %src2)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> %passthru
ret <16 x i8> %3
}
define <16 x i8> @test_vgf2p8mulb_128_maskz(<16 x i8> %src1, <16 x i8> %src2, i16 %mask) {
; X86-LABEL: test_vgf2p8mulb_128_maskz:
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_128_maskz:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8> %src1, <16 x i8> %src2)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
%4 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> %passthru
%5 = xor <16 x i8> %3, %4
%6 = xor <16 x i8> %5, %2
ret <16 x i8> %6
ret <16 x i8> %3
}
declare <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8>, <32 x i8>)
define <32 x i8> @test_vgf2p8mulb_256(<32 x i8> %src1, <32 x i8> %src2, <32 x i8> %passthru, i32 %mask) {
; X86-LABEL: test_vgf2p8mulb_256:
define <32 x i8> @test_vgf2p8mulb_256(<32 x i8> %src1, <32 x i8> %src2) {
; CHECK-LABEL: test_vgf2p8mulb_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8> %src1, <32 x i8> %src2)
ret <32 x i8> %1
}
define <32 x i8> @test_vgf2p8mulb_256_mask(<32 x i8> %src1, <32 x i8> %src2, <32 x i8> %passthru, i32 %mask) {
; X86-LABEL: test_vgf2p8mulb_256_mask:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xd9]
; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xe1]
; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xcf,0xd1]
; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_256:
; X64-LABEL: test_vgf2p8mulb_256_mask:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xd9]
; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xe1]
; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xcf,0xd1]
; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8> %src1, <32 x i8> %src2)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> %passthru
ret <32 x i8> %3
}
define <32 x i8> @test_vgf2p8mulb_256_maskz(<32 x i8> %src1, <32 x i8> %src2, i32 %mask) {
; X86-LABEL: test_vgf2p8mulb_256_maskz:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_256_maskz:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8> %src1, <32 x i8> %src2)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
%4 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> %passthru
%5 = xor <32 x i8> %3, %4
%6 = xor <32 x i8> %5, %2
ret <32 x i8> %6
ret <32 x i8> %3
}
declare <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8>, <64 x i8>)
define <64 x i8> @test_vgf2p8mulb_512(<64 x i8> %src1, <64 x i8> %src2, <64 x i8> %passthru, i64 %mask) {
; X86-LABEL: test_vgf2p8mulb_512:
define <64 x i8> @test_vgf2p8mulb_512(<64 x i8> %src1, <64 x i8> %src2) {
; CHECK-LABEL: test_vgf2p8mulb_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8> %src1, <64 x i8> %src2)
ret <64 x i8> %1
}
define <64 x i8> @test_vgf2p8mulb_512_mask(<64 x i8> %src1, <64 x i8> %src2, <64 x i8> %passthru, i64 %mask) {
; X86-LABEL: test_vgf2p8mulb_512_mask:
; X86: # %bb.0:
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xd9]
; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xe1]
; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xcf,0xd1]
; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_512:
; X64-LABEL: test_vgf2p8mulb_512_mask:
; X64: # %bb.0:
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xd9]
; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xe1]
; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xcf,0xd1]
; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8> %src1, <64 x i8> %src2)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> %passthru
ret <64 x i8> %3
}
define <64 x i8> @test_vgf2p8mulb_512_maskz(<64 x i8> %src1, <64 x i8> %src2, i64 %mask) {
; X86-LABEL: test_vgf2p8mulb_512_maskz:
; X86: # %bb.0:
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_vgf2p8mulb_512_maskz:
; X64: # %bb.0:
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8> %src1, <64 x i8> %src2)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
%4 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> %passthru
%5 = xor <64 x i8> %3, %4
%6 = xor <64 x i8> %5, %2
ret <64 x i8> %6
ret <64 x i8> %3
}

768
llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
View File

File diff suppressed because it is too large Load Diff

86
llvm/test/CodeGen/X86/avx512-intrinsics.ll
View File

@ -4980,112 +4980,98 @@ define <4 x float>@test_int_x86_avx512_mask_cvt_sd2ss_round(<4 x float> %x0,<2 x
declare <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i32)
define <16 x i32>@test_int_x86_avx512_pternlog_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_d_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i32 33)
ret <16 x i32> %1
}
define <16 x i32>@test_int_x86_avx512_mask_pternlog_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x4) {
; X64-LABEL: test_int_x86_avx512_mask_pternlog_d_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3
; X64-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm3
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm0 {%k1}
; X64-NEXT: vpaddd %zmm3, %zmm0, %zmm0
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_mask_pternlog_d_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3
; X86-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
; X86-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm0 {%k1}
; X86-NEXT: vpaddd %zmm3, %zmm0, %zmm0
; X86-NEXT: retl
%1 = call <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i32 33)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i32 33)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
ret <16 x i32> %3
}
define <16 x i32>@test_int_x86_avx512_maskz_pternlog_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x4) {
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_d_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3
; X64-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm3
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X64-NEXT: vpaddd %zmm3, %zmm0, %zmm0
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_d_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3
; X86-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
; X86-NEXT: vpternlogd $33, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X86-NEXT: vpaddd %zmm3, %zmm0, %zmm0
; X86-NEXT: retl
%1 = call <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i32 33)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> zeroinitializer
%4 = call <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i32 33)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
ret <16 x i32> %3
}
declare <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i32)
define <8 x i64>@test_int_x86_avx512_pternlog_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_q_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i32 33)
ret <8 x i64> %1
}
define <8 x i64>@test_int_x86_avx512_mask_pternlog_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x4) {
; X64-LABEL: test_int_x86_avx512_mask_pternlog_q_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3
; X64-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm3
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm0 {%k1}
; X64-NEXT: vpaddq %zmm3, %zmm0, %zmm0
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_mask_pternlog_q_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3
; X86-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm3
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm0 {%k1}
; X86-NEXT: vpaddq %zmm3, %zmm0, %zmm0
; X86-NEXT: retl
%1 = call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x0
%4 = call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i32 33)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
ret <8 x i64> %3
}
define <8 x i64>@test_int_x86_avx512_maskz_pternlog_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x4) {
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_q_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3
; X64-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm3
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X64-NEXT: vpaddq %zmm3, %zmm0, %zmm0
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_q_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3
; X86-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm3
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpternlogq $33, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X86-NEXT: vpaddq %zmm3, %zmm0, %zmm0
; X86-NEXT: retl
%1 = call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> zeroinitializer
%4 = call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i32 33)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
ret <8 x i64> %3
}
define i32 @test_x86_avx512_comi_sd_eq_sae(<2 x double> %a0, <2 x double> %a1) {
@ -5603,13 +5589,13 @@ define <16 x float>@test_int_x86_avx512_maskz_fixupimm_ps_512(<16 x float> %x0,
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vmovaps %zmm0, %zmm3
; X64-NEXT: vfixupimmps $5, %zmm2, %zmm1, %zmm3
; X64-NEXT: vmovaps %zmm0, %zmm4
; X64-NEXT: vfixupimmps $5, %zmm2, %zmm1, %zmm4 {%k1} {z}
; X64-NEXT: vxorps %xmm2, %xmm2, %xmm2
; X64-NEXT: vfixupimmps $5, {sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X64-NEXT: vaddps %zmm0, %zmm4, %zmm0
; X64-NEXT: vaddps %zmm3, %zmm0, %zmm0
; X64-NEXT: vfixupimmps $5, %zmm2, %zmm1, %zmm3 {%k1} {z}
; X64-NEXT: vxorps %xmm4, %xmm4, %xmm4
; X64-NEXT: vmovaps %zmm0, %zmm5
; X64-NEXT: vfixupimmps $6, {sae}, %zmm4, %zmm1, %zmm5 {%k1} {z}
; X64-NEXT: vaddps %zmm5, %zmm3, %zmm3
; X64-NEXT: vfixupimmps $7, %zmm2, %zmm1, %zmm0
; X64-NEXT: vaddps %zmm0, %zmm3, %zmm0
; X64-NEXT: retq
;
; X86-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_512:
@ -5617,16 +5603,16 @@ define <16 x float>@test_int_x86_avx512_maskz_fixupimm_ps_512(<16 x float> %x0,
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
; X86-NEXT: vmovaps %zmm0, %zmm3
; X86-NEXT: vfixupimmps $5, %zmm2, %zmm1, %zmm3 {%k1} {z}
; X86-NEXT: vmovaps %zmm0, %zmm4
; X86-NEXT: vfixupimmps $5, %zmm2, %zmm1, %zmm4
; X86-NEXT: vxorps %xmm2, %xmm2, %xmm2
; X86-NEXT: vfixupimmps $5, {sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
; X86-NEXT: vxorps %xmm4, %xmm4, %xmm4
; X86-NEXT: vmovaps %zmm0, %zmm5
; X86-NEXT: vfixupimmps $6, {sae}, %zmm4, %zmm1, %zmm5 {%k1} {z}
; X86-NEXT: vaddps %zmm5, %zmm3, %zmm3
; X86-NEXT: vfixupimmps $7, %zmm2, %zmm1, %zmm0
; X86-NEXT: vaddps %zmm0, %zmm3, %zmm0
; X86-NEXT: vaddps %zmm4, %zmm0, %zmm0
; X86-NEXT: retl
%res = call <16 x float> @llvm.x86.avx512.maskz.fixupimm.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x i32> %x2, i32 5, i16 %x4, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.fixupimm.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x i32> zeroinitializer, i32 5, i16 %x4, i32 8)
%res2 = call <16 x float> @llvm.x86.avx512.maskz.fixupimm.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x i32> %x2, i32 5, i16 -1, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.fixupimm.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x i32> zeroinitializer, i32 6, i16 %x4, i32 8)
%res2 = call <16 x float> @llvm.x86.avx512.maskz.fixupimm.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x i32> %x2, i32 7, i16 -1, i32 4)
%res3 = fadd <16 x float> %res, %res1
%res4 = fadd <16 x float> %res3, %res2
ret <16 x float> %res4

64
llvm/test/CodeGen/X86/avx512-rotate.ll
View File

@ -15,26 +15,26 @@ declare <16 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.512(<16 x i32>, <16 x i32
define <16 x i32> @test_splat_rol_v16i32(<16 x i32> %x0, <16 x i32> %x1, i16 %x2) {
; KNL-LABEL: test_splat_rol_v16i32:
; KNL: # %bb.0:
; KNL-NEXT: vprold $5, %zmm0, %zmm2
; KNL-NEXT: kmovw %edi, %k1
; KNL-NEXT: vprold $5, %zmm0, %zmm1 {%k1}
; KNL-NEXT: vprold $5, %zmm0, %zmm0 {%k1} {z}
; KNL-NEXT: vprold $6, %zmm0, %zmm2 {%k1} {z}
; KNL-NEXT: vpaddd %zmm2, %zmm1, %zmm1
; KNL-NEXT: vprold $7, %zmm0, %zmm0
; KNL-NEXT: vpaddd %zmm0, %zmm1, %zmm0
; KNL-NEXT: vpaddd %zmm2, %zmm0, %zmm0
; KNL-NEXT: retq
;
; SKX-LABEL: test_splat_rol_v16i32:
; SKX: # %bb.0:
; SKX-NEXT: vprold $5, %zmm0, %zmm2
; SKX-NEXT: kmovd %edi, %k1
; SKX-NEXT: vprold $5, %zmm0, %zmm1 {%k1}
; SKX-NEXT: vprold $5, %zmm0, %zmm0 {%k1} {z}
; SKX-NEXT: vprold $6, %zmm0, %zmm2 {%k1} {z}
; SKX-NEXT: vpaddd %zmm2, %zmm1, %zmm1
; SKX-NEXT: vprold $7, %zmm0, %zmm0
; SKX-NEXT: vpaddd %zmm0, %zmm1, %zmm0
; SKX-NEXT: vpaddd %zmm2, %zmm0, %zmm0
; SKX-NEXT: retq
%res = call <16 x i32> @llvm.x86.avx512.mask.prolv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> %x1, i16 %x2)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.prolv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> zeroinitializer, i16 %x2)
%res2 = call <16 x i32> @llvm.x86.avx512.mask.prolv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> %x1, i16 -1)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.prolv.d.512(<16 x i32> %x0, <16 x i32> <i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6>, <16 x i32> zeroinitializer, i16 %x2)
%res2 = call <16 x i32> @llvm.x86.avx512.mask.prolv.d.512(<16 x i32> %x0, <16 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>, <16 x i32> %x1, i16 -1)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res3, %res2
ret <16 x i32> %res4
@ -43,26 +43,26 @@ define <16 x i32> @test_splat_rol_v16i32(<16 x i32> %x0, <16 x i32> %x1, i16 %x2
define <8 x i64>@test_splat_rol_v8i64(<8 x i64> %x0, <8 x i64> %x1, i8 %x2) {
; KNL-LABEL: test_splat_rol_v8i64:
; KNL: # %bb.0:
; KNL-NEXT: vprolq $5, %zmm0, %zmm2
; KNL-NEXT: kmovw %edi, %k1
; KNL-NEXT: vprolq $5, %zmm0, %zmm1 {%k1}
; KNL-NEXT: vprolq $5, %zmm0, %zmm0 {%k1} {z}
; KNL-NEXT: vprolq $6, %zmm0, %zmm2 {%k1} {z}
; KNL-NEXT: vpaddq %zmm2, %zmm1, %zmm1
; KNL-NEXT: vprolq $7, %zmm0, %zmm0
; KNL-NEXT: vpaddq %zmm0, %zmm1, %zmm0
; KNL-NEXT: vpaddq %zmm2, %zmm0, %zmm0
; KNL-NEXT: retq
;
; SKX-LABEL: test_splat_rol_v8i64:
; SKX: # %bb.0:
; SKX-NEXT: vprolq $5, %zmm0, %zmm2
; SKX-NEXT: kmovd %edi, %k1
; SKX-NEXT: vprolq $5, %zmm0, %zmm1 {%k1}
; SKX-NEXT: vprolq $5, %zmm0, %zmm0 {%k1} {z}
; SKX-NEXT: vprolq $6, %zmm0, %zmm2 {%k1} {z}
; SKX-NEXT: vpaddq %zmm2, %zmm1, %zmm1
; SKX-NEXT: vprolq $7, %zmm0, %zmm0
; SKX-NEXT: vpaddq %zmm0, %zmm1, %zmm0
; SKX-NEXT: vpaddq %zmm2, %zmm0, %zmm0
; SKX-NEXT: retq
%res = call <8 x i64> @llvm.x86.avx512.mask.prolv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> %x1, i8 %x2)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.prolv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> zeroinitializer, i8 %x2)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.prolv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> %x1, i8 -1)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.prolv.q.512(<8 x i64> %x0, <8 x i64> <i64 6, i64 6, i64 6, i64 6, i64 6, i64 6, i64 6, i64 6>, <8 x i64> zeroinitializer, i8 %x2)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.prolv.q.512(<8 x i64> %x0, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>, <8 x i64> %x1, i8 -1)
%res3 = add <8 x i64> %res, %res1
%res4 = add <8 x i64> %res3, %res2
ret <8 x i64> %res4
@ -71,26 +71,26 @@ define <8 x i64>@test_splat_rol_v8i64(<8 x i64> %x0, <8 x i64> %x1, i8 %x2) {
define <16 x i32> @test_splat_ror_v16i32(<16 x i32> %x0, <16 x i32> %x1, i16 %x2) {
; KNL-LABEL: test_splat_ror_v16i32:
; KNL: # %bb.0:
; KNL-NEXT: vprord $5, %zmm0, %zmm2
; KNL-NEXT: kmovw %edi, %k1
; KNL-NEXT: vprord $5, %zmm0, %zmm1 {%k1}
; KNL-NEXT: vprord $5, %zmm0, %zmm0 {%k1} {z}
; KNL-NEXT: vprord $6, %zmm0, %zmm2 {%k1} {z}
; KNL-NEXT: vpaddd %zmm2, %zmm1, %zmm1
; KNL-NEXT: vprord $7, %zmm0, %zmm0
; KNL-NEXT: vpaddd %zmm0, %zmm1, %zmm0
; KNL-NEXT: vpaddd %zmm2, %zmm0, %zmm0
; KNL-NEXT: retq
;
; SKX-LABEL: test_splat_ror_v16i32:
; SKX: # %bb.0:
; SKX-NEXT: vprord $5, %zmm0, %zmm2
; SKX-NEXT: kmovd %edi, %k1
; SKX-NEXT: vprord $5, %zmm0, %zmm1 {%k1}
; SKX-NEXT: vprord $5, %zmm0, %zmm0 {%k1} {z}
; SKX-NEXT: vprord $6, %zmm0, %zmm2 {%k1} {z}
; SKX-NEXT: vpaddd %zmm2, %zmm1, %zmm1
; SKX-NEXT: vprord $7, %zmm0, %zmm0
; SKX-NEXT: vpaddd %zmm0, %zmm1, %zmm0
; SKX-NEXT: vpaddd %zmm2, %zmm0, %zmm0
; SKX-NEXT: retq
%res = call <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> %x1, i16 %x2)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> zeroinitializer, i16 %x2)
%res2 = call <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32> %x0, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>, <16 x i32> %x1, i16 -1)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32> %x0, <16 x i32> <i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6, i32 6>, <16 x i32> zeroinitializer, i16 %x2)
%res2 = call <16 x i32> @llvm.x86.avx512.mask.prorv.d.512(<16 x i32> %x0, <16 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>, <16 x i32> %x1, i16 -1)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res3, %res2
ret <16 x i32> %res4
@ -99,26 +99,26 @@ define <16 x i32> @test_splat_ror_v16i32(<16 x i32> %x0, <16 x i32> %x1, i16 %x2
define <8 x i64>@test_splat_ror_v8i64(<8 x i64> %x0, <8 x i64> %x1, i8 %x2) {
; KNL-LABEL: test_splat_ror_v8i64:
; KNL: # %bb.0:
; KNL-NEXT: vprorq $5, %zmm0, %zmm2
; KNL-NEXT: kmovw %edi, %k1
; KNL-NEXT: vprorq $5, %zmm0, %zmm1 {%k1}
; KNL-NEXT: vprorq $5, %zmm0, %zmm0 {%k1} {z}
; KNL-NEXT: vprorq $6, %zmm0, %zmm2 {%k1} {z}
; KNL-NEXT: vpaddq %zmm2, %zmm1, %zmm1
; KNL-NEXT: vprorq $7, %zmm0, %zmm0
; KNL-NEXT: vpaddq %zmm0, %zmm1, %zmm0
; KNL-NEXT: vpaddq %zmm2, %zmm0, %zmm0
; KNL-NEXT: retq
;
; SKX-LABEL: test_splat_ror_v8i64:
; SKX: # %bb.0:
; SKX-NEXT: vprorq $5, %zmm0, %zmm2
; SKX-NEXT: kmovd %edi, %k1
; SKX-NEXT: vprorq $5, %zmm0, %zmm1 {%k1}
; SKX-NEXT: vprorq $5, %zmm0, %zmm0 {%k1} {z}
; SKX-NEXT: vprorq $6, %zmm0, %zmm2 {%k1} {z}
; SKX-NEXT: vpaddq %zmm2, %zmm1, %zmm1
; SKX-NEXT: vprorq $7, %zmm0, %zmm0
; SKX-NEXT: vpaddq %zmm0, %zmm1, %zmm0
; SKX-NEXT: vpaddq %zmm2, %zmm0, %zmm0
; SKX-NEXT: retq
%res = call <8 x i64> @llvm.x86.avx512.mask.prorv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> %x1, i8 %x2)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.prorv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> zeroinitializer, i8 %x2)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.prorv.q.512(<8 x i64> %x0, <8 x i64> <i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5, i64 5>, <8 x i64> %x1, i8 -1)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.prorv.q.512(<8 x i64> %x0, <8 x i64> <i64 6, i64 6, i64 6, i64 6, i64 6, i64 6, i64 6, i64 6>, <8 x i64> zeroinitializer, i8 %x2)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.prorv.q.512(<8 x i64> %x0, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>, <8 x i64> %x1, i8 -1)
%res3 = add <8 x i64> %res, %res1
%res4 = add <8 x i64> %res3, %res2
ret <8 x i64> %res4

409
llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll
View File

@ -1081,32 +1081,47 @@ define <32 x i16>@test_int_x86_avx512_maskz_pmovsxb_w_512(<32 x i8> %x0, i32 %x2
declare <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16>, <8 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psrl_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psrl_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsrlw %xmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xd1,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psrl_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrl_w_512:
; X86: # %bb.0:
; X86-NEXT: vpsrlw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xd1,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xd1,0xd1]
; X86-NEXT: vpsrlw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xd1,0xc1]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrl_w_512:
; X64: # %bb.0:
; X64-NEXT: vpsrlw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xd1,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xd1,0xd1]
; X64-NEXT: vpsrlw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xd1,0xc1]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psrl_w_512(<32 x i16> %x0, <8 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrl_w_512:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xd1,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrl_w_512:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xd1,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrl.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16>, i32, <32 x i16>, i32)
@ -1114,26 +1129,26 @@ declare <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16>, i32, <32 x i16>
define <32 x i16>@test_int_x86_avx512_mask_psrl_wi_512(<32 x i16> %x0, i32 %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrl_wi_512:
; X86: # %bb.0:
; X86-NEXT: vpsrlw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xd0,0x03]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vpsrlw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xd0,0x03]
; X86-NEXT: vpsrlw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xd0,0x03]
; X86-NEXT: vpsrlw $4, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xd0,0x04]
; X86-NEXT: vpsrlw $5, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xd0,0x05]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrl_wi_512:
; X64: # %bb.0:
; X64-NEXT: vpsrlw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xd0,0x03]
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vpsrlw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xd0,0x03]
; X64-NEXT: vpsrlw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xd0,0x03]
; X64-NEXT: vpsrlw $4, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xd0,0x04]
; X64-NEXT: vpsrlw $5, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xd0,0x05]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 -1)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 3, <32 x i16> zeroinitializer, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 4, <32 x i16> %x2, i32 -1)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psrl.wi.512(<32 x i16> %x0, i32 5, <32 x i16> zeroinitializer, i32 %x3)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
@ -1141,32 +1156,47 @@ define <32 x i16>@test_int_x86_avx512_mask_psrl_wi_512(<32 x i16> %x0, i32 %x1,
declare <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16>, <8 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psra_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psra_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsraw %xmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe1,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psra_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psra_w_512:
; X86: # %bb.0:
; X86-NEXT: vpsraw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe1,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsraw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe1,0xd1]
; X86-NEXT: vpsraw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xe1,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psra_w_512:
; X64: # %bb.0:
; X64-NEXT: vpsraw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe1,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsraw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe1,0xd1]
; X64-NEXT: vpsraw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xe1,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psra_w_512(<32 x i16> %x0, <8 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psra_w_512:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsraw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xe1,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psra_w_512:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsraw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xe1,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psra.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16>, i32, <32 x i16>, i32)
@ -1174,26 +1204,26 @@ declare <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16>, i32, <32 x i16>
define <32 x i16>@test_int_x86_avx512_mask_psra_wi_512(<32 x i16> %x0, i32 %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psra_wi_512:
; X86: # %bb.0:
; X86-NEXT: vpsraw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xe0,0x03]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vpsraw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xe0,0x03]
; X86-NEXT: vpsraw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xe0,0x03]
; X86-NEXT: vpaddw %zmm2, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc2]
; X86-NEXT: vpsraw $4, %zmm0, %zmm2 {%k1} {z} # encoding: [0x62,0xf1,0x6d,0xc9,0x71,0xe0,0x04]
; X86-NEXT: vpsraw $5, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0x71,0xe0,0x05]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psra_wi_512:
; X64: # %bb.0:
; X64-NEXT: vpsraw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xe0,0x03]
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vpsraw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xe0,0x03]
; X64-NEXT: vpsraw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xe0,0x03]
; X64-NEXT: vpaddw %zmm2, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc2]
; X64-NEXT: vpsraw $4, %zmm0, %zmm2 {%k1} {z} # encoding: [0x62,0xf1,0x6d,0xc9,0x71,0xe0,0x04]
; X64-NEXT: vpsraw $5, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0x71,0xe0,0x05]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16> %x0, i32 3, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 -1)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16> %x0, i32 4, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psra.wi.512(<32 x i16> %x0, i32 5, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
@ -1201,32 +1231,47 @@ define <32 x i16>@test_int_x86_avx512_mask_psra_wi_512(<32 x i16> %x0, i32 %x1,
declare <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16>, <8 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psll_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psll_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsllw %xmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xf1,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psll_w_512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psll_w_512:
; X86: # %bb.0:
; X86-NEXT: vpsllw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xf1,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xf1,0xd1]
; X86-NEXT: vpsllw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xf1,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psll_w_512:
; X64: # %bb.0:
; X64-NEXT: vpsllw %xmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xf1,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllw %xmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xf1,0xd1]
; X64-NEXT: vpsllw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xf1,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psll_w_512(<32 x i16> %x0, <8 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psll_w_512:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xf1,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psll_w_512:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllw %xmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0xf1,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psll.w.512(<32 x i16> %x0, <8 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16>, i32, <32 x i16>, i32)
@ -1234,26 +1279,26 @@ declare <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16>, i32, <32 x i16>
define <32 x i16>@test_int_x86_avx512_mask_psll_wi_512(<32 x i16> %x0, i32 %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psll_wi_512:
; X86: # %bb.0:
; X86-NEXT: vpsllw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xf0,0x03]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vpsllw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xf0,0x03]
; X86-NEXT: vpsllw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xf0,0x03]
; X86-NEXT: vpaddw %zmm2, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc2]
; X86-NEXT: vpsllw $4, %zmm0, %zmm2 {%k1} {z} # encoding: [0x62,0xf1,0x6d,0xc9,0x71,0xf0,0x04]
; X86-NEXT: vpsllw $5, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0x71,0xf0,0x05]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psll_wi_512:
; X64: # %bb.0:
; X64-NEXT: vpsllw $3, %zmm0, %zmm2 # encoding: [0x62,0xf1,0x6d,0x48,0x71,0xf0,0x03]
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vpsllw $3, %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf1,0x75,0x49,0x71,0xf0,0x03]
; X64-NEXT: vpsllw $3, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf1,0x7d,0xc9,0x71,0xf0,0x03]
; X64-NEXT: vpaddw %zmm2, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc2]
; X64-NEXT: vpsllw $4, %zmm0, %zmm2 {%k1} {z} # encoding: [0x62,0xf1,0x6d,0xc9,0x71,0xf0,0x04]
; X64-NEXT: vpsllw $5, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0x71,0xf0,0x05]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16> %x0, i32 3, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16> %x0, i32 3, <32 x i16> %x2, i32 -1)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16> %x0, i32 4, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psll.wi.512(<32 x i16> %x0, i32 5, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
@ -2541,98 +2586,117 @@ declare i32 @llvm.x86.avx512.mask.ucmp.w.512(<32 x i16>, <32 x i16>, i32, i32) n
declare <64 x i8> @llvm.x86.avx512.mask.pavg.b.512(<64 x i8>, <64 x i8>, <64 x i8>, i64)
define <64 x i8>@mm512_avg_epu8(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 %x3) {
; X86-LABEL: mm512_avg_epu8:
; X86: # %bb.0:
; X86-NEXT: vpavgb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xd9]
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X86-NEXT: vpaddb %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfc,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: mm512_avg_epu8:
; X64: # %bb.0:
; X64-NEXT: vpavgb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xd9]
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X64-NEXT: vpaddb %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfc,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <64 x i8> @llvm.x86.avx512.mask.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 %x3)
%res1 = call <64 x i8> @llvm.x86.avx512.mask.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 -1)
%res2 = add <64 x i8> %res, %res1
ret <64 x i8> %res2
; CHECK-LABEL: mm512_avg_epu8:
; CHECK: # %bb.0:
; CHECK-NEXT: vpavgb %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <64 x i8> @llvm.x86.avx512.mask.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 -1)
ret <64 x i8> %res
}
define <64 x i8>@mm512_mask_avg_epu8(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 %x3) {
; X86-LABEL: mm512_mask_avg_epu8:
; X86: # %bb.0:
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: mm512_mask_avg_epu8:
; X64: # %bb.0:
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <64 x i8> @llvm.x86.avx512.mask.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 %x3)
ret <64 x i8> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.pavg.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@mm512_avg_epu16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: mm512_avg_epu16:
; CHECK-LABEL: mm512_avg_epu16:
; CHECK: # %bb.0:
; CHECK-NEXT: vpavgw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@mm512_mask_avg_epu16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: mm512_mask_avg_epu16:
; X86: # %bb.0:
; X86-NEXT: vpavgw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe3,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: mm512_avg_epu16:
; X64-LABEL: mm512_mask_avg_epu16:
; X64: # %bb.0:
; X64-NEXT: vpavgw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe3,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res2
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.pabs.w.512(<32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_pabs_w_512(<32 x i16> %x0, <32 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_pabs_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpabsw %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x48,0x1d,0xc0]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pabs.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_pabs_w_512(<32 x i16> %x0, <32 x i16> %x1, i32 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_pabs_w_512:
; X86: # %bb.0:
; X86-NEXT: vpabsw %zmm0, %zmm2 # encoding: [0x62,0xf2,0x7d,0x48,0x1d,0xd0]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpabsw %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x1d,0xc8]
; X86-NEXT: vpaddw %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc2]
; X86-NEXT: vmovdqa64 %zmm1, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pabs_w_512:
; X64: # %bb.0:
; X64-NEXT: vpabsw %zmm0, %zmm2 # encoding: [0x62,0xf2,0x7d,0x48,0x1d,0xd0]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpabsw %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x1d,0xc8]
; X64-NEXT: vpaddw %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfd,0xc2]
; X64-NEXT: vmovdqa64 %zmm1, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pabs.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 %x2)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.pabs.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 -1)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res2
ret <32 x i16> %res
}
declare <64 x i8> @llvm.x86.avx512.mask.pabs.b.512(<64 x i8>, <64 x i8>, i64)
define <64 x i8>@test_int_x86_avx512_pabs_b_512(<64 x i8> %x0, <64 x i8> %x1) {
; CHECK-LABEL: test_int_x86_avx512_pabs_b_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpabsb %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x48,0x1c,0xc0]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <64 x i8> @llvm.x86.avx512.mask.pabs.b.512(<64 x i8> %x0, <64 x i8> %x1, i64 -1)
ret <64 x i8> %res
}
define <64 x i8>@test_int_x86_avx512_mask_pabs_b_512(<64 x i8> %x0, <64 x i8> %x1, i64 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_pabs_b_512:
; X86: # %bb.0:
; X86-NEXT: vpabsb %zmm0, %zmm2 # encoding: [0x62,0xf2,0x7d,0x48,0x1c,0xd0]
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpabsb %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x1c,0xc8]
; X86-NEXT: vpaddb %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfc,0xc2]
; X86-NEXT: vmovdqa64 %zmm1, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pabs_b_512:
; X64: # %bb.0:
; X64-NEXT: vpabsb %zmm0, %zmm2 # encoding: [0x62,0xf2,0x7d,0x48,0x1c,0xd0]
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vpabsb %zmm0, %zmm1 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x1c,0xc8]
; X64-NEXT: vpaddb %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf1,0x75,0x48,0xfc,0xc2]
; X64-NEXT: vmovdqa64 %zmm1, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <64 x i8> @llvm.x86.avx512.mask.pabs.b.512(<64 x i8> %x0, <64 x i8> %x1, i64 %x2)
%res1 = call <64 x i8> @llvm.x86.avx512.mask.pabs.b.512(<64 x i8> %x0, <64 x i8> %x1, i64 -1)
%res2 = add <64 x i8> %res, %res1
ret <64 x i8> %res2
ret <64 x i8> %res
}
declare i64 @llvm.x86.avx512.ptestm.b.512(<64 x i8>, <64 x i8>, i64)
@ -3069,26 +3133,26 @@ declare <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8>, <64 x i8>, i32,
define <32 x i16>@test_int_x86_avx512_mask_dbpsadbw_512(<64 x i8> %x0, <64 x i8> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_dbpsadbw_512:
; X86: # %bb.0:
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xd9,0x02]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x42,0xd1,0x02]
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xc1,0x02]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vdbpsadbw $3, %zmm1, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xd9,0x03]
; X86-NEXT: vdbpsadbw $4, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xc1,0x04]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_dbpsadbw_512:
; X64: # %bb.0:
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xd9,0x02]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x42,0xd1,0x02]
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xc1,0x02]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vdbpsadbw $3, %zmm1, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xd9,0x03]
; X64-NEXT: vdbpsadbw $4, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xc1,0x04]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2, <32 x i16> %x3, i32 %x4)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2, <32 x i16> zeroinitializer, i32 %x4)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2, <32 x i16> %x3, i32 -1)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 3, <32 x i16> zeroinitializer, i32 %x4)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 4, <32 x i16> %x3, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
@ -4102,92 +4166,137 @@ declare <64 x i8> @llvm.x86.avx512.mask.psubs.b.512(<64 x i8>, <64 x i8>, <64 x
declare <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psrlv32hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrlv32hi:
; X86: # %bb.0:
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x10,0xd1]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrlv32hi:
; X64: # %bb.0:
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x10,0xd1]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrlv32hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrlv32hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psrav32_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsravw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrav32_hi:
; X86: # %bb.0:
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x11,0xd1]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrav32_hi:
; X64: # %bb.0:
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x11,0xd1]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrav32_hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrav32_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psrav32.hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psllv32hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psllv32hi:
; X86: # %bb.0:
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x12,0xd1]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psllv32hi:
; X64: # %bb.0:
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x12,0xd1]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
%res2 = call <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res3, %res2
ret <32 x i16> %res4
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_maskz_psllv32hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psllv32hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psllv32hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> zeroinitializer, i32 %x3)
ret <32 x i16> %res
}
declare <32 x i8> @llvm.x86.avx512.mask.pmov.wb.512(<32 x i16>, <32 x i8>, i32)

284
llvm/test/CodeGen/X86/avx512bw-intrinsics.ll
View File

@ -763,54 +763,64 @@ define <32 x i16>@test_int_x86_avx512_mask_vpermi2var_hi_512(<32 x i16> %x0, <32
declare <64 x i8> @llvm.x86.avx512.pavg.b.512(<64 x i8>, <64 x i8>)
define <64 x i8> @test_int_x86_avx512_pavg_b_512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pavg_b_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpavgb %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <64 x i8> @llvm.x86.avx512.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1)
ret <64 x i8> %1
}
define <64 x i8> @test_int_x86_avx512_mask_pavg_b_512(<64 x i8> %x0, <64 x i8> %x1, <64 x i8> %x2, i64 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_b_512:
; X86: # %bb.0:
; X86-NEXT: vpavgb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xd9]
; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X86-NEXT: vpaddb %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfc,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_b_512:
; X64: # %bb.0:
; X64-NEXT: vpavgb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe0,0xd9]
; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
; X64-NEXT: vpavgb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe0,0xd1]
; X64-NEXT: vpaddb %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfc,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <64 x i8> @llvm.x86.avx512.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1)
%2 = bitcast i64 %x3 to <64 x i1>
%3 = select <64 x i1> %2, <64 x i8> %1, <64 x i8> %x2
%4 = call <64 x i8> @llvm.x86.avx512.pavg.b.512(<64 x i8> %x0, <64 x i8> %x1)
%res2 = add <64 x i8> %3, %4
ret <64 x i8> %res2
ret <64 x i8> %3
}
declare <32 x i16> @llvm.x86.avx512.pavg.w.512(<32 x i16>, <32 x i16>)
define <32 x i16> @test_int_x86_avx512_pavg_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pavg_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpavgw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16> @test_int_x86_avx512_mask_pavg_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_w_512:
; X86: # %bb.0:
; X86-NEXT: vpavgw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe3,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_w_512:
; X64: # %bb.0:
; X64-NEXT: vpavgw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe3,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe3,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.pavg.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
ret <32 x i16> %3
}
declare <64 x i8> @llvm.x86.avx512.pshuf.b.512(<64 x i8>, <64 x i8>)
@ -864,80 +874,95 @@ define <64 x i8>@test_int_x86_avx512_pshuf_b_512_maskz(<64 x i8> %x0, <64 x i8>
declare <32 x i16> @llvm.x86.avx512.pmulhu.w.512(<32 x i16>, <32 x i16>)
define <32 x i16> @test_int_x86_avx512_pmulhu_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pmulhu_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmulhuw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe4,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmulhu.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16> @test_int_x86_avx512_mask_pmulhu_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhu_w_512:
; X86: # %bb.0:
; X86-NEXT: vpmulhuw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe4,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhuw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe4,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhu_w_512:
; X64: # %bb.0:
; X64-NEXT: vpmulhuw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe4,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhuw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe4,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmulhu.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.pmulhu.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
ret <32 x i16> %3
}
declare <32 x i16> @llvm.x86.avx512.pmulh.w.512(<32 x i16>, <32 x i16>)
define <32 x i16> @test_int_x86_avx512_pmulh_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pmulh_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmulhw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xe5,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmulh.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16> @test_int_x86_avx512_mask_pmulh_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulh_w_512:
; X86: # %bb.0:
; X86-NEXT: vpmulhw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe5,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe5,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulh_w_512:
; X64: # %bb.0:
; X64-NEXT: vpmulhw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xe5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xe5,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmulh.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.pmulh.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
ret <32 x i16> %3
}
declare <32 x i16> @llvm.x86.avx512.pmul.hr.sw.512(<32 x i16>, <32 x i16>)
define <32 x i16> @test_int_x86_avx512_pmulhr_sw_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pmulhr_sw_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmulhrsw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x48,0x0b,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmul.hr.sw.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16> @test_int_x86_avx512_mask_pmulhr_sw_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhr_sw_512:
; X86: # %bb.0:
; X86-NEXT: vpmulhrsw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0x0b,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhrsw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x0b,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhr_sw_512:
; X64: # %bb.0:
; X64-NEXT: vpmulhrsw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0x0b,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhrsw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x0b,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmul.hr.sw.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.pmul.hr.sw.512(<32 x i16> %x0, <32 x i16> %x1)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
ret <32 x i16> %3
}
define <32 x i8>@test_int_x86_avx512_pmov_wb_512(<32 x i16> %x0) {
@ -1151,54 +1176,64 @@ define void @test_int_x86_avx512_mask_pmovus_wb_mem_512(i8* %ptr, <32 x i16> %x1
declare <32 x i16> @llvm.x86.avx512.pmaddubs.w.512(<64 x i8>, <64 x i8>)
define <32 x i16> @test_int_x86_avx512_pmaddubs_w_512(<64 x i8> %x0, <64 x i8> %x1) {
; CHECK-LABEL: test_int_x86_avx512_pmaddubs_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmaddubsw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0x7d,0x48,0x04,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmaddubs.w.512(<64 x i8> %x0, <64 x i8> %x1)
ret <32 x i16> %1
}
define <32 x i16> @test_int_x86_avx512_mask_pmaddubs_w_512(<64 x i8> %x0, <64 x i8> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddubs_w_512:
; X86: # %bb.0:
; X86-NEXT: vpmaddubsw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0x04,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmaddubsw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x04,0xd1]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddubs_w_512:
; X64: # %bb.0:
; X64-NEXT: vpmaddubsw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0x04,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddubsw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0x04,0xd1]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.pmaddubs.w.512(<64 x i8> %x0, <64 x i8> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.pmaddubs.w.512(<64 x i8> %x0, <64 x i8> %x1)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
ret <32 x i16> %3
}
declare <16 x i32> @llvm.x86.avx512.pmaddw.d.512(<32 x i16>, <32 x i16>)
define <16 x i32> @test_int_x86_avx512_pmaddw_d_512(<32 x i16> %x0, <32 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_pmaddw_d_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmaddwd %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xf5,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.pmaddw.d.512(<32 x i16> %x0, <32 x i16> %x1)
ret <16 x i32> %1
}
define <16 x i32> @test_int_x86_avx512_mask_pmaddw_d_512(<32 x i16> %x0, <32 x i16> %x1, <16 x i32> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddw_d_512:
; X86: # %bb.0:
; X86-NEXT: vpmaddwd %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xf5,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmaddwd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xf5,0xd1]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddw_d_512:
; X64: # %bb.0:
; X64-NEXT: vpmaddwd %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7d,0x48,0xf5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddwd %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x49,0xf5,0xd1]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.pmaddw.d.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x2
%4 = call <16 x i32> @llvm.x86.avx512.pmaddw.d.512(<32 x i16> %x0, <32 x i16> %x1)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
ret <16 x i32> %3
}
declare <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8>, <64 x i8>, i32)
@ -1206,30 +1241,30 @@ declare <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8>, <64 x i8>, i32)
define <32 x i16>@test_int_x86_avx512_mask_dbpsadbw_512(<64 x i8> %x0, <64 x i8> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_dbpsadbw_512:
; X86: # %bb.0:
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xd9,0x02]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x42,0xd1,0x02]
; X86-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xc1,0x02]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vdbpsadbw $3, %zmm1, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xd9,0x03]
; X86-NEXT: vdbpsadbw $4, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xc1,0x04]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_dbpsadbw_512:
; X64: # %bb.0:
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xd9,0x02]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x42,0xd1,0x02]
; X64-NEXT: vdbpsadbw $2, %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xc1,0x02]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vdbpsadbw $3, %zmm1, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xc9,0x42,0xd9,0x03]
; X64-NEXT: vdbpsadbw $4, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x42,0xc1,0x04]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2)
%2 = bitcast i32 %x4 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x3
%4 = call <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2)
%4 = call <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 3)
%5 = bitcast i32 %x4 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%7 = call <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 2)
%7 = call <32 x i16> @llvm.x86.avx512.dbpsadbw.512(<64 x i8> %x0, <64 x i8> %x1, i32 4)
%res3 = add <32 x i16> %3, %6
%res4 = add <32 x i16> %res3, %7
ret <32 x i16> %res4
@ -1274,70 +1309,100 @@ define <32 x i16> @test_x86_avx512_psrlv_w_512_const() optsize {
ret <32 x i16> %res1
}
define <32 x i16>@test_int_x86_avx512_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psrlv32hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrlv.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16>@test_int_x86_avx512_mask_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrlv32hi:
; X86: # %bb.0:
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x10,0xd1]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrlv32hi:
; X64: # %bb.0:
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x10,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x10,0xd1]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrlv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.psrlv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%5 = bitcast i32 %x3 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%7 = call <32 x i16> @llvm.x86.avx512.psrlv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res3 = add <32 x i16> %3, %6
%res4 = add <32 x i16> %res3, %7
ret <32 x i16> %res4
ret <32 x i16> %3
}
define <32 x i16>@test_int_x86_avx512_maskz_psrlv32hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrlv32hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrlv32hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x10,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrlv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> zeroinitializer
ret <32 x i16> %3
}
declare <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16>, <32 x i16>)
define <32 x i16>@test_int_x86_avx512_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psrav32_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsravw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16>@test_int_x86_avx512_mask_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrav32_hi:
; X86: # %bb.0:
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x11,0xd1]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrav32_hi:
; X64: # %bb.0:
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x11,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x11,0xd1]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16> %x0, <32 x i16> %x1)
%5 = bitcast i32 %x3 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%7 = call <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res3 = add <32 x i16> %3, %6
%res4 = add <32 x i16> %res3, %7
ret <32 x i16> %res4
ret <32 x i16> %3
}
define <32 x i16>@test_int_x86_avx512_maskz_psrav32_hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrav32_hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrav32_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x11,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psrav.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> zeroinitializer
ret <32 x i16> %3
}
define <32 x i16>@test_int_x86_avx512_mask_psrav32_hi_const(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
@ -1362,36 +1427,51 @@ define <32 x i16>@test_int_x86_avx512_mask_psrav32_hi_const(<32 x i16> %x0, <32
ret <32 x i16> %1
}
define <32 x i16>@test_int_x86_avx512_psllv32hi(<32 x i16> %x0, <32 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psllv32hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psllv.w.512(<32 x i16> %x0, <32 x i16> %x1)
ret <32 x i16> %1
}
define <32 x i16>@test_int_x86_avx512_mask_psllv32hi(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psllv32hi:
; X86: # %bb.0:
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x12,0xd1]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X86-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psllv32hi:
; X64: # %bb.0:
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0xfd,0x48,0x12,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0x12,0xd1]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X64-NEXT: vpaddw %zmm3, %zmm0, %zmm0 # encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc3]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: vmovdqa64 %zmm2, %zmm0 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psllv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x2
%4 = call <32 x i16> @llvm.x86.avx512.psllv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%5 = bitcast i32 %x3 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%7 = call <32 x i16> @llvm.x86.avx512.psllv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%res3 = add <32 x i16> %3, %6
%res4 = add <32 x i16> %res3, %7
ret <32 x i16> %res4
ret <32 x i16> %3
}
define <32 x i16>@test_int_x86_avx512_maskz_psllv32hi(<32 x i16> %x0, <32 x i16> %x1, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psllv32hi:
; X86: # %bb.0:
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psllv32hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %zmm1, %zmm0, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0x12,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.psllv.w.512(<32 x i16> %x0, <32 x i16> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> zeroinitializer
ret <32 x i16> %3
}
declare <32 x i16> @llvm.x86.avx512.permvar.hi.512(<32 x i16>, <32 x i16>)

771
llvm/test/CodeGen/X86/avx512bwvl-intrinsics-upgrade.ll
View File

File diff suppressed because it is too large Load Diff

437
llvm/test/CodeGen/X86/avx512bwvl-intrinsics.ll
View File

@ -1291,25 +1291,21 @@ declare <16 x i8> @llvm.x86.sse2.pavg.b(<16 x i8>, <16 x i8>)
define <16 x i8> @test_int_x86_avx512_mask_pavg_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_b_128:
; X86: # %bb.0:
; X86-NEXT: vpavgb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe0,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe0,0xd1]
; X86-NEXT: vpaddb %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfc,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_b_128:
; X64: # %bb.0:
; X64-NEXT: vpavgb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe0,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe0,0xd1]
; X64-NEXT: vpaddb %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfc,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i8> @llvm.x86.sse2.pavg.b(<16 x i8> %x0, <16 x i8> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i8> %1, <16 x i8> %x2
%4 = call <16 x i8> @llvm.x86.sse2.pavg.b(<16 x i8> %x0, <16 x i8> %x1)
%res2 = add <16 x i8> %3, %4
ret <16 x i8> %res2
ret <16 x i8> %3
}
declare <32 x i8> @llvm.x86.avx2.pavg.b(<32 x i8>, <32 x i8>)
@ -1317,25 +1313,21 @@ declare <32 x i8> @llvm.x86.avx2.pavg.b(<32 x i8>, <32 x i8>)
define <32 x i8> @test_int_x86_avx512_mask_pavg_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_b_256:
; X86: # %bb.0:
; X86-NEXT: vpavgb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe0,0xd9]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe0,0xd1]
; X86-NEXT: vpaddb %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfc,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_b_256:
; X64: # %bb.0:
; X64-NEXT: vpavgb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe0,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe0,0xd1]
; X64-NEXT: vpaddb %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfc,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i8> @llvm.x86.avx2.pavg.b(<32 x i8> %x0, <32 x i8> %x1)
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i8> %1, <32 x i8> %x2
%4 = call <32 x i8> @llvm.x86.avx2.pavg.b(<32 x i8> %x0, <32 x i8> %x1)
%res2 = add <32 x i8> %3, %4
ret <32 x i8> %res2
ret <32 x i8> %3
}
declare <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16>, <8 x i16>)
@ -1343,26 +1335,22 @@ declare <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16>, <8 x i16>)
define <8 x i16> @test_int_x86_avx512_mask_pavg_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_w_128:
; X86: # %bb.0:
; X86-NEXT: vpavgw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe3,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpavgw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe3,0xd1]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_w_128:
; X64: # %bb.0:
; X64-NEXT: vpavgw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe3,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe3,0xd1]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16> %x0, <8 x i16> %x1)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
ret <8 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.pavg.w(<16 x i16>, <16 x i16>)
@ -1370,25 +1358,21 @@ declare <16 x i16> @llvm.x86.avx2.pavg.w(<16 x i16>, <16 x i16>)
define <16 x i16> @test_int_x86_avx512_mask_pavg_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pavg_w_256:
; X86: # %bb.0:
; X86-NEXT: vpavgw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe3,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpavgw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe3,0xd1]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pavg_w_256:
; X64: # %bb.0:
; X64-NEXT: vpavgw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe3,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpavgw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe3,0xd1]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx2.pavg.w(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx2.pavg.w(<16 x i16> %x0, <16 x i16> %x1)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
ret <16 x i16> %3
}
declare <8 x i16> @llvm.x86.sse2.pmulhu.w(<8 x i16>, <8 x i16>)
@ -1396,26 +1380,22 @@ declare <8 x i16> @llvm.x86.sse2.pmulhu.w(<8 x i16>, <8 x i16>)
define <8 x i16> @test_int_x86_avx512_mask_pmulhu_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhu_w_128:
; X86: # %bb.0:
; X86-NEXT: vpmulhuw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe4,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmulhuw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe4,0xd1]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhu_w_128:
; X64: # %bb.0:
; X64-NEXT: vpmulhuw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe4,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhuw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe4,0xd1]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.sse2.pmulhu.w(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.sse2.pmulhu.w(<8 x i16> %x0, <8 x i16> %x1)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
ret <8 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.pmulhu.w(<16 x i16>, <16 x i16>)
@ -1423,25 +1403,21 @@ declare <16 x i16> @llvm.x86.avx2.pmulhu.w(<16 x i16>, <16 x i16>)
define <16 x i16> @test_int_x86_avx512_mask_pmulhu_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhu_w_256:
; X86: # %bb.0:
; X86-NEXT: vpmulhuw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe4,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhuw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe4,0xd1]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhu_w_256:
; X64: # %bb.0:
; X64-NEXT: vpmulhuw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe4,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhuw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe4,0xd1]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx2.pmulhu.w(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx2.pmulhu.w(<16 x i16> %x0, <16 x i16> %x1)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
ret <16 x i16> %3
}
declare <8 x i16> @llvm.x86.sse2.pmulh.w(<8 x i16>, <8 x i16>)
@ -1449,26 +1425,22 @@ declare <8 x i16> @llvm.x86.sse2.pmulh.w(<8 x i16>, <8 x i16>)
define <8 x i16> @test_int_x86_avx512_mask_pmulh_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulh_w_128:
; X86: # %bb.0:
; X86-NEXT: vpmulhw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe5,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmulhw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe5,0xd1]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulh_w_128:
; X64: # %bb.0:
; X64-NEXT: vpmulhw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xe5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xe5,0xd1]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.sse2.pmulh.w(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.sse2.pmulh.w(<8 x i16> %x0, <8 x i16> %x1)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
ret <8 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.pmulh.w(<16 x i16>, <16 x i16>)
@ -1476,25 +1448,21 @@ declare <16 x i16> @llvm.x86.avx2.pmulh.w(<16 x i16>, <16 x i16>)
define <16 x i16> @test_int_x86_avx512_mask_pmulh_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulh_w_256:
; X86: # %bb.0:
; X86-NEXT: vpmulhw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe5,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe5,0xd1]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulh_w_256:
; X64: # %bb.0:
; X64-NEXT: vpmulhw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xe5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xe5,0xd1]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx2.pmulh.w(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx2.pmulh.w(<16 x i16> %x0, <16 x i16> %x1)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
ret <16 x i16> %3
}
declare <8 x i16> @llvm.x86.ssse3.pmul.hr.sw.128(<8 x i16>, <8 x i16>)
@ -1502,26 +1470,22 @@ declare <8 x i16> @llvm.x86.ssse3.pmul.hr.sw.128(<8 x i16>, <8 x i16>)
define <8 x i16> @test_int_x86_avx512_mask_pmulhr_sw_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhr_sw_128:
; X86: # %bb.0:
; X86-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x0b,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x0b,0xd1]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhr_sw_128:
; X64: # %bb.0:
; X64-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x0b,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x0b,0xd1]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.ssse3.pmul.hr.sw.128(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.ssse3.pmul.hr.sw.128(<8 x i16> %x0, <8 x i16> %x1)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
ret <8 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.pmul.hr.sw(<16 x i16>, <16 x i16>)
@ -1529,25 +1493,21 @@ declare <16 x i16> @llvm.x86.avx2.pmul.hr.sw(<16 x i16>, <16 x i16>)
define <16 x i16> @test_int_x86_avx512_mask_pmulhr_sw_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmulhr_sw_256:
; X86: # %bb.0:
; X86-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x0b,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x0b,0xd1]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmulhr_sw_256:
; X64: # %bb.0:
; X64-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x0b,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x0b,0xd1]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx2.pmul.hr.sw(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx2.pmul.hr.sw(<16 x i16> %x0, <16 x i16> %x1)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
ret <16 x i16> %3
}
declare <16 x i8> @llvm.x86.avx512.mask.pmov.wb.128(<8 x i16>, <16 x i8>, i8)
@ -1941,27 +1901,23 @@ declare <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16>, <8 x i16>)
define <4 x i32> @test_int_x86_avx512_mask_pmaddw_d_128(<8 x i16> %x0, <8 x i16> %x1, <4 x i32> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddw_d_128:
; X86: # %bb.0:
; X86-NEXT: vpmaddwd %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xf5,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmaddwd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xf5,0xd1]
; X86-NEXT: vpaddd %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddw_d_128:
; X64: # %bb.0:
; X64-NEXT: vpmaddwd %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xf5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddwd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x09,0xf5,0xd1]
; X64-NEXT: vpaddd %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x2
%4 = call <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16> %x0, <8 x i16> %x1)
%res2 = add <4 x i32> %3, %4
ret <4 x i32> %res2
ret <4 x i32> %3
}
declare <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16>, <16 x i16>)
@ -1969,26 +1925,22 @@ declare <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16>, <16 x i16>)
define <8 x i32> @test_int_x86_avx512_mask_pmaddw_d_256(<16 x i16> %x0, <16 x i16> %x1, <8 x i32> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddw_d_256:
; X86: # %bb.0:
; X86-NEXT: vpmaddwd %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xf5,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmaddwd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xf5,0xd1]
; X86-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddw_d_256:
; X64: # %bb.0:
; X64-NEXT: vpmaddwd %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xf5,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddwd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf1,0x7d,0x29,0xf5,0xd1]
; X64-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x2
%4 = call <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16> %x0, <16 x i16> %x1)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
ret <8 x i32> %3
}
declare <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8>, <16 x i8>)
@ -1996,26 +1948,22 @@ declare <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8>, <16 x i8>)
define <8 x i16> @test_int_x86_avx512_mask_pmaddubs_w_128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddubs_w_128:
; X86: # %bb.0:
; X86-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x04,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x04,0xd1]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddubs_w_128:
; X64: # %bb.0:
; X64-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x04,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x04,0xd1]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8> %x0, <16 x i8> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8> %x0, <16 x i8> %x1)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
ret <8 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8>, <32 x i8>)
@ -2023,25 +1971,21 @@ declare <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8>, <32 x i8>)
define <16 x i16> @test_int_x86_avx512_mask_pmaddubs_w_256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_pmaddubs_w_256:
; X86: # %bb.0:
; X86-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x04,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x04,0xd1]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pmaddubs_w_256:
; X64: # %bb.0:
; X64-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x04,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x04,0xd1]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8> %x0, <32 x i8> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8> %x0, <32 x i8> %x1)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
ret <16 x i16> %3
}
declare <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8>, <16 x i8>, i32)
@ -2049,31 +1993,31 @@ declare <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8>, <16 x i8>, i32)
define <8 x i16>@test_int_x86_avx512_mask_dbpsadbw_128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_dbpsadbw_128:
; X86: # %bb.0:
; X86-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x42,0xd9,0x02]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x42,0xd1,0x02]
; X86-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x42,0xc1,0x02]
; X86-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X86-NEXT: vdbpsadbw $3, %xmm1, %xmm0, %xmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x42,0xd9,0x03]
; X86-NEXT: vdbpsadbw $4, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x42,0xc1,0x04]
; X86-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_dbpsadbw_128:
; X64: # %bb.0:
; X64-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x42,0xd9,0x02]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x42,0xd1,0x02]
; X64-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x42,0xc1,0x02]
; X64-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X64-NEXT: vdbpsadbw $3, %xmm1, %xmm0, %xmm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x42,0xd9,0x03]
; X64-NEXT: vdbpsadbw $4, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x42,0xc1,0x04]
; X64-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x3
%4 = call <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2)
%4 = call <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 3)
%5 = bitcast i8 %x4 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i16> %4, <8 x i16> zeroinitializer
%7 = call <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2)
%7 = call <8 x i16> @llvm.x86.avx512.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 4)
%res3 = add <8 x i16> %3, %6
%res4 = add <8 x i16> %7, %res3
ret <8 x i16> %res4
@ -2084,30 +2028,30 @@ declare <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8>, <32 x i8>, i32)
define <16 x i16>@test_int_x86_avx512_mask_dbpsadbw_256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_dbpsadbw_256:
; X86: # %bb.0:
; X86-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x42,0xd9,0x02]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x42,0xd1,0x02]
; X86-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xa9,0x42,0xc1,0x02]
; X86-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X86-NEXT: vdbpsadbw $3, %ymm1, %ymm0, %ymm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xa9,0x42,0xd9,0x03]
; X86-NEXT: vdbpsadbw $4, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x42,0xc1,0x04]
; X86-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_dbpsadbw_256:
; X64: # %bb.0:
; X64-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x42,0xd9,0x02]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x42,0xd1,0x02]
; X64-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xa9,0x42,0xc1,0x02]
; X64-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X64-NEXT: vdbpsadbw $3, %ymm1, %ymm0, %ymm3 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0xa9,0x42,0xd9,0x03]
; X64-NEXT: vdbpsadbw $4, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x42,0xc1,0x04]
; X64-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x3
%4 = call <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2)
%4 = call <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 3)
%5 = bitcast i16 %x4 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i16> %4, <16 x i16> zeroinitializer
%7 = call <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2)
%7 = call <16 x i16> @llvm.x86.avx512.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 4)
%res3 = add <16 x i16> %3, %6
%res4 = add <16 x i16> %res3, %7
ret <16 x i16> %res4
@ -2115,63 +2059,94 @@ define <16 x i16>@test_int_x86_avx512_mask_dbpsadbw_256(<32 x i8> %x0, <32 x i8>
declare <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_psrlv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psrlv16_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x10,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
ret <16 x i16> %res
}
define <16 x i16>@test_int_x86_avx512_mask_psrlv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrlv16_hi:
; X86: # %bb.0:
; X86-NEXT: vpsrlvw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x10,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x10,0xd1]
; X86-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x10,0xc1]
; X86-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrlv16_hi:
; X64: # %bb.0:
; X64-NEXT: vpsrlvw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x10,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x10,0xd1]
; X64-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x10,0xc1]
; X64-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res3 = add <16 x i16> %res, %res1
%res4 = add <16 x i16> %res3, %res2
ret <16 x i16> %res4
ret <16 x i16> %res
}
define <16 x i16>@test_int_x86_avx512_maskz_psrlv16_hi(<16 x i16> %x0, <16 x i16> %x1, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrlv16_hi:
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x10,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrlv16_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x10,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
ret <16 x i16> %res
}
declare <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_psrlv8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_psrlv8_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x10,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
ret <8 x i16> %res
}
define <8 x i16>@test_int_x86_avx512_mask_psrlv8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrlv8_hi:
; X86: # %bb.0:
; X86-NEXT: vpsrlvw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x10,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsrlvw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x10,0xd1]
; X86-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x10,0xc1]
; X86-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrlv8_hi:
; X64: # %bb.0:
; X64-NEXT: vpsrlvw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x10,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x10,0xd1]
; X64-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x10,0xc1]
; X64-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res3 = add <8 x i16> %res, %res1
%res4 = add <8 x i16> %res3, %res2
ret <8 x i16> %res4
ret <8 x i16> %res
}
define <8 x i16>@test_int_x86_avx512_maskz_psrlv8_hi(<8 x i16> %x0, <8 x i16> %x1, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrlv8_hi:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x10,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrlv8_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x10,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
ret <8 x i16> %res
}
@ -2225,136 +2200,198 @@ declare <16 x i16> @llvm.x86.avx512.psrlv.w.256(<16 x i16>, <16 x i16>)
declare <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16>, <16 x i16>)
define <16 x i16>@test_int_x86_avx512_psrav16_hi(<16 x i16> %x0, <16 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psrav16_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x11,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %x0, <16 x i16> %x1)
ret <16 x i16> %1
}
define <16 x i16>@test_int_x86_avx512_mask_psrav16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrav16_hi:
; X86: # %bb.0:
; X86-NEXT: vpsravw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x11,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x11,0xd1]
; X86-NEXT: vpsravw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x11,0xc1]
; X86-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrav16_hi:
; X64: # %bb.0:
; X64-NEXT: vpsravw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x11,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x11,0xd1]
; X64-NEXT: vpsravw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x11,0xc1]
; X64-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %x0, <16 x i16> %x1)
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i16> %4, <16 x i16> zeroinitializer
%7 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %x0, <16 x i16> %x1)
%res3 = add <16 x i16> %3, %6
%res4 = add <16 x i16> %res3, %7
ret <16 x i16> %res4
ret <16 x i16> %3
}
define <16 x i16>@test_int_x86_avx512_maskz_psrav16_hi(<16 x i16> %x0, <16 x i16> %x1, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrav16_hi:
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsravw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x11,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrav16_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x11,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> zeroinitializer
ret <16 x i16> %3
}
declare <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16>, <8 x i16>)
define <8 x i16>@test_int_x86_avx512_psrav8_hi(<8 x i16> %x0, <8 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psrav8_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x11,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %x0, <8 x i16> %x1)
ret <8 x i16> %1
}
define <8 x i16>@test_int_x86_avx512_mask_psrav8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psrav8_hi:
; X86: # %bb.0:
; X86-NEXT: vpsravw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x11,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsravw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x11,0xd1]
; X86-NEXT: vpsravw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x11,0xc1]
; X86-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psrav8_hi:
; X64: # %bb.0:
; X64-NEXT: vpsravw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x11,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x11,0xd1]
; X64-NEXT: vpsravw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x11,0xc1]
; X64-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %x0, <8 x i16> %x1)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i16> %4, <8 x i16> zeroinitializer
%7 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %x0, <8 x i16> %x1)
%res3 = add <8 x i16> %3, %6
%res4 = add <8 x i16> %res3, %7
ret <8 x i16> %res4
ret <8 x i16> %3
}
define <8 x i16>@test_int_x86_avx512_maskz_psrav8_hi(<8 x i16> %x0, <8 x i16> %x1, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psrav8_hi:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsravw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x11,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psrav8_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsravw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x11,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> zeroinitializer
ret <8 x i16> %3
}
define <16 x i16>@test_int_x86_avx512_psllv16_hi(<16 x i16> %x0, <16 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psllv16_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x12,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %x0, <16 x i16> %x1)
ret <16 x i16> %1
}
define <16 x i16>@test_int_x86_avx512_mask_psllv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psllv16_hi:
; X86: # %bb.0:
; X86-NEXT: vpsllvw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x12,0xd9]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x12,0xd1]
; X86-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x12,0xc1]
; X86-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psllv16_hi:
; X64: # %bb.0:
; X64-NEXT: vpsllvw %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf2,0xfd,0x28,0x12,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x12,0xd1]
; X64-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x12,0xc1]
; X64-NEXT: vpaddw %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfd,0xc3]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: vmovdqa %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x2
%4 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %x0, <16 x i16> %x1)
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i16> %4, <16 x i16> zeroinitializer
%7 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %x0, <16 x i16> %x1)
%res3 = add <16 x i16> %3, %6
%res4 = add <16 x i16> %res3, %7
ret <16 x i16> %res4
ret <16 x i16> %3
}
define <16 x i16>@test_int_x86_avx512_maskz_psllv16_hi(<16 x i16> %x0, <16 x i16> %x1, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psllv16_hi:
; X86: # %bb.0:
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x12,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psllv16_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xa9,0x12,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %x0, <16 x i16> %x1)
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> zeroinitializer
ret <16 x i16> %3
}
define <8 x i16>@test_int_x86_avx512_psllv8_hi(<8 x i16> %x0, <8 x i16> %x1) {
; CHECK-LABEL: test_int_x86_avx512_psllv8_hi:
; CHECK: # %bb.0:
; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x12,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %x0, <8 x i16> %x1)
ret <8 x i16> %1
}
define <8 x i16>@test_int_x86_avx512_mask_psllv8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_psllv8_hi:
; X86: # %bb.0:
; X86-NEXT: vpsllvw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x12,0xd9]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsllvw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x12,0xd1]
; X86-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x12,0xc1]
; X86-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_psllv8_hi:
; X64: # %bb.0:
; X64-NEXT: vpsllvw %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf2,0xfd,0x08,0x12,0xd9]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x12,0xd1]
; X64-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x12,0xc1]
; X64-NEXT: vpaddw %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfd,0xc3]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: vmovdqa %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x2
%4 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %x0, <8 x i16> %x1)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i16> %4, <8 x i16> zeroinitializer
%7 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %x0, <8 x i16> %x1)
%res3 = add <8 x i16> %3, %6
%res4 = add <8 x i16> %res3, %7
ret <8 x i16> %res4
ret <8 x i16> %3
}
define <8 x i16>@test_int_x86_avx512_maskz_psllv8_hi(<8 x i16> %x0, <8 x i16> %x1, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_psllv8_hi:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x12,0xc1]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_psllv8_hi:
; X64: # %bb.0:
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0x89,0x12,0xc1]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %x0, <8 x i16> %x1)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> zeroinitializer
ret <8 x i16> %3
}
define <8 x i16> @test_int_x86_avx512_psllv_w_128_const() optsize {

267
llvm/test/CodeGen/X86/avx512cdvl-intrinsics-upgrade.ll
View File

@ -4,108 +4,139 @@
declare <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_vplzcnt_d_128(<4 x i32> %x0, <4 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vplzcnt_d_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vplzcntd %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_mask_vplzcnt_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_d_128:
; X86: # %bb.0:
; X86-NEXT: vplzcntd %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %xmm0, %xmm1 {%k1}
; X86-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_d_128:
; X64: # %bb.0:
; X64-NEXT: vplzcntd %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %xmm0, %xmm1 {%k1}
; X64-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%res = call <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
%res3 = call <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %x2)
%res2 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_maskz_vplzcnt_d_128(<4 x i32> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vplzcnt_d_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vplzcnt_d_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%res = call <4 x i32> @llvm.x86.avx512.mask.lzcnt.d.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %x2)
ret <4 x i32> %res
}
declare <8 x i32> @llvm.x86.avx512.mask.lzcnt.d.256(<8 x i32>, <8 x i32>, i8)
define <8 x i32>@test_int_x86_avx512_vplzcnt_d_256(<8 x i32> %x0, <8 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vplzcnt_d_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vplzcntd %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <8 x i32> @llvm.x86.avx512.mask.lzcnt.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
ret <8 x i32> %res
}
define <8 x i32>@test_int_x86_avx512_mask_vplzcnt_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_d_256:
; X86: # %bb.0:
; X86-NEXT: vplzcntd %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %ymm0, %ymm1 {%k1}
; X86-NEXT: vpaddd %ymm2, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_d_256:
; X64: # %bb.0:
; X64-NEXT: vplzcntd %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %ymm0, %ymm1 {%k1}
; X64-NEXT: vpaddd %ymm2, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%res = call <8 x i32> @llvm.x86.avx512.mask.lzcnt.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.lzcnt.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
%res2 = add <8 x i32> %res, %res1
ret <8 x i32> %res2
ret <8 x i32> %res
}
declare <2 x i64> @llvm.x86.avx512.mask.lzcnt.q.128(<2 x i64>, <2 x i64>, i8)
define <2 x i64>@test_int_x86_avx512_vplzcnt_q_128(<2 x i64> %x0, <2 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vplzcnt_q_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vplzcntq %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <2 x i64> @llvm.x86.avx512.mask.lzcnt.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
ret <2 x i64> %res
}
define <2 x i64>@test_int_x86_avx512_mask_vplzcnt_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_q_128:
; X86: # %bb.0:
; X86-NEXT: vplzcntq %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntq %xmm0, %xmm1 {%k1}
; X86-NEXT: vpaddq %xmm2, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_q_128:
; X64: # %bb.0:
; X64-NEXT: vplzcntq %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntq %xmm0, %xmm1 {%k1}
; X64-NEXT: vpaddq %xmm2, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%res = call <2 x i64> @llvm.x86.avx512.mask.lzcnt.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.lzcnt.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
ret <2 x i64> %res2
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.mask.lzcnt.q.256(<4 x i64>, <4 x i64>, i8)
define <4 x i64>@test_int_x86_avx512_vplzcnt_q_256(<4 x i64> %x0, <4 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vplzcnt_q_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vplzcntq %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <4 x i64> @llvm.x86.avx512.mask.lzcnt.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
ret <4 x i64> %res
}
define <4 x i64>@test_int_x86_avx512_mask_vplzcnt_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_q_256:
; X86: # %bb.0:
; X86-NEXT: vplzcntq %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntq %ymm0, %ymm1 {%k1}
; X86-NEXT: vpaddq %ymm2, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_q_256:
; X64: # %bb.0:
; X64-NEXT: vplzcntq %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntq %ymm0, %ymm1 {%k1}
; X64-NEXT: vpaddq %ymm2, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%res = call <4 x i64> @llvm.x86.avx512.mask.lzcnt.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.lzcnt.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
ret <4 x i64> %res2
ret <4 x i64> %res
}
define <8 x i32> @test_x86_vbroadcastmw_256(i16 %a0) {
@ -182,125 +213,189 @@ declare <2 x i64> @llvm.x86.avx512.broadcastmb.128(i8)
declare <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32>, <4 x i32>, i8)
define <4 x i32> @test_int_x86_avx512_vpconflict_d_128(<4 x i32> %x0, <4 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_d_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictd %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
ret <4 x i32> %res
}
define <4 x i32> @test_int_x86_avx512_mask_vpconflict_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_d_128:
; X86: # %bb.0:
; X86-NEXT: vpconflictd %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %xmm0, %xmm1 {%k1}
; X86-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_d_128:
; X64: # %bb.0:
; X64-NEXT: vpconflictd %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %xmm0, %xmm1 {%k1}
; X64-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%res = call <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
%res3 = call <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %x2)
%res2 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
ret <4 x i32> %res
}
define <4 x i32> @test_int_x86_avx512_maskz_vpconflict_d_128(<4 x i32> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_d_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_d_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%res = call <4 x i32> @llvm.x86.avx512.mask.conflict.d.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %x2)
ret <4 x i32> %res
}
declare <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32>, <8 x i32>, i8)
define <8 x i32> @test_int_x86_avx512_vpconflict_d_256(<8 x i32> %x0, <8 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_d_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictd %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
ret <8 x i32> %res
}
define <8 x i32> @test_int_x86_avx512_mask_vpconflict_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_d_256:
; X86: # %bb.0:
; X86-NEXT: vpconflictd %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %ymm0, %ymm1 {%k1}
; X86-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: vpaddd %ymm0, %ymm2, %ymm0
; X86-NEXT: vpaddd %ymm0, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_d_256:
; X64: # %bb.0:
; X64-NEXT: vpconflictd %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %ymm0, %ymm1 {%k1}
; X64-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: vpaddd %ymm0, %ymm2, %ymm0
; X64-NEXT: vpaddd %ymm0, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%res = call <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
%res2 = call <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32> %x0, <8 x i32> zeroinitializer, i8 %x2)
%res3 = add <8 x i32> %res, %res1
%res4 = add <8 x i32> %res2, %res3
ret <8 x i32> %res4
ret <8 x i32> %res
}
define <8 x i32> @test_int_x86_avx512_maskz_vpconflict_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_d_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_d_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: retq
%res = call <8 x i32> @llvm.x86.avx512.mask.conflict.d.256(<8 x i32> %x0, <8 x i32> zeroinitializer, i8 %x2)
ret <8 x i32> %res
}
declare <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64>, <2 x i64>, i8)
define <2 x i64> @test_int_x86_avx512_vpconflict_q_128(<2 x i64> %x0, <2 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_q_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictq %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
ret <2 x i64> %res
}
define <2 x i64> @test_int_x86_avx512_mask_vpconflict_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_q_128:
; X86: # %bb.0:
; X86-NEXT: vpconflictq %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %xmm0, %xmm1 {%k1}
; X86-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddq %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_q_128:
; X64: # %bb.0:
; X64-NEXT: vpconflictq %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %xmm0, %xmm1 {%k1}
; X64-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddq %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%res = call <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
%res2 = call <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64> %x0, <2 x i64> zeroinitializer, i8 %x2)
%res3 = add <2 x i64> %res, %res1
%res4 = add <2 x i64> %res2, %res3
ret <2 x i64> %res4
ret <2 x i64> %res
}
define <2 x i64> @test_int_x86_avx512_maskz_vpconflict_q_128(<2 x i64> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_q_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_q_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%res = call <2 x i64> @llvm.x86.avx512.mask.conflict.q.128(<2 x i64> %x0, <2 x i64> zeroinitializer, i8 %x2)
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64>, <4 x i64>, i8)
define <4 x i64> @test_int_x86_avx512_vpconflict_q_256(<4 x i64> %x0, <4 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_q_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictq %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%res = call <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
ret <4 x i64> %res
}
define <4 x i64> @test_int_x86_avx512_mask_vpconflict_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_q_256:
; X86: # %bb.0:
; X86-NEXT: vpconflictq %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %ymm0, %ymm1 {%k1}
; X86-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0
; X86-NEXT: vpaddq %ymm0, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_q_256:
; X64: # %bb.0:
; X64-NEXT: vpconflictq %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %ymm0, %ymm1 {%k1}
; X64-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0
; X64-NEXT: vpaddq %ymm0, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%res = call <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64> %x0, <4 x i64> zeroinitializer, i8 %x2)
%res3 = add <4 x i64> %res, %res1
%res4 = add <4 x i64> %res2, %res3
ret <4 x i64> %res4
ret <4 x i64> %res
}
define <4 x i64> @test_int_x86_avx512_maskz_vpconflict_q_256(<4 x i64> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_q_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_q_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: retq
%res = call <4 x i64> @llvm.x86.avx512.mask.conflict.q.256(<4 x i64> %x0, <4 x i64> zeroinitializer, i8 %x2)
ret <4 x i64> %res
}

259
llvm/test/CodeGen/X86/avx512cdvl-intrinsics.ll
View File

@ -5,255 +5,314 @@
define <4 x i32> @test_int_x86_avx512_mask_vplzcnt_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_d_128:
; X86: # %bb.0:
; X86-NEXT: vplzcntd %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %xmm0, %xmm1 {%k1}
; X86-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_d_128:
; X64: # %bb.0:
; X64-NEXT: vplzcntd %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %xmm0, %xmm1 {%k1}
; X64-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%1 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %x0, i1 false)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> %x1
%4 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %x0, i1 false)
%5 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %x0, i1 false)
%6 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract, <4 x i32> %5, <4 x i32> zeroinitializer
%res2 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %res2, %7
ret <4 x i32> %res4
ret <4 x i32> %3
}
define <4 x i32> @test_int_x86_avx512_maskz_vplzcnt_d_128(<4 x i32> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vplzcnt_d_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vplzcnt_d_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%1 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %x0, i1 false)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> zeroinitializer
ret <4 x i32> %3
}
declare <4 x i32> @llvm.ctlz.v4i32(<4 x i32>, i1) #0
define <8 x i32> @test_int_x86_avx512_mask_vplzcnt_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_d_256:
; X86: # %bb.0:
; X86-NEXT: vplzcntd %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntd %ymm0, %ymm1 {%k1}
; X86-NEXT: vpaddd %ymm2, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_d_256:
; X64: # %bb.0:
; X64-NEXT: vplzcntd %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntd %ymm0, %ymm1 {%k1}
; X64-NEXT: vpaddd %ymm2, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%1 = call <8 x i32> @llvm.ctlz.v8i32(<8 x i32> %x0, i1 false)
%2 = bitcast i8 %x2 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x1
%4 = call <8 x i32> @llvm.ctlz.v8i32(<8 x i32> %x0, i1 false)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
ret <8 x i32> %3
}
declare <8 x i32> @llvm.ctlz.v8i32(<8 x i32>, i1) #0
define <2 x i64> @test_int_x86_avx512_mask_vplzcnt_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_q_128:
; X86: # %bb.0:
; X86-NEXT: vplzcntq %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntq %xmm0, %xmm1 {%k1}
; X86-NEXT: vpaddq %xmm2, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_q_128:
; X64: # %bb.0:
; X64-NEXT: vplzcntq %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntq %xmm0, %xmm1 {%k1}
; X64-NEXT: vpaddq %xmm2, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%1 = call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %x0, i1 false)
%2 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x1
%4 = call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %x0, i1 false)
%res2 = add <2 x i64> %3, %4
ret <2 x i64> %res2
ret <2 x i64> %3
}
declare <2 x i64> @llvm.ctlz.v2i64(<2 x i64>, i1) #0
define <4 x i64> @test_int_x86_avx512_mask_vplzcnt_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vplzcnt_q_256:
; X86: # %bb.0:
; X86-NEXT: vplzcntq %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vplzcntq %ymm0, %ymm1 {%k1}
; X86-NEXT: vpaddq %ymm2, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vplzcnt_q_256:
; X64: # %bb.0:
; X64-NEXT: vplzcntq %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vplzcntq %ymm0, %ymm1 {%k1}
; X64-NEXT: vpaddq %ymm2, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%1 = call <4 x i64> @llvm.ctlz.v4i64(<4 x i64> %x0, i1 false)
%2 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x1
%4 = call <4 x i64> @llvm.ctlz.v4i64(<4 x i64> %x0, i1 false)
%res2 = add <4 x i64> %3, %4
ret <4 x i64> %res2
ret <4 x i64> %3
}
declare <4 x i64> @llvm.ctlz.v4i64(<4 x i64>, i1) #0
define <4 x i32> @test_int_x86_avx512_vpconflict_d_128(<4 x i32> %x0, <4 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_d_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictd %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %x0)
ret <4 x i32> %1
}
define <4 x i32> @test_int_x86_avx512_mask_vpconflict_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_d_128:
; X86: # %bb.0:
; X86-NEXT: vpconflictd %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %xmm0, %xmm1 {%k1}
; X86-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_d_128:
; X64: # %bb.0:
; X64-NEXT: vpconflictd %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %xmm0, %xmm1 {%k1}
; X64-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddd %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%1 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> %x1
%4 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %x0)
%5 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %x0)
%6 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract, <4 x i32> %5, <4 x i32> zeroinitializer
%res2 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %res2, %7
ret <4 x i32> %res4
ret <4 x i32> %3
}
define <4 x i32> @test_int_x86_avx512_maskz_vpconflict_d_128(<4 x i32> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_d_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_d_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%1 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> zeroinitializer
ret <4 x i32> %3
}
define <8 x i32> @test_int_x86_avx512_vpconflict_d_256(<8 x i32> %x0, <8 x i32> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_d_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictd %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %x0)
ret <8 x i32> %1
}
define <8 x i32> @test_int_x86_avx512_mask_vpconflict_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_d_256:
; X86: # %bb.0:
; X86-NEXT: vpconflictd %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %ymm0, %ymm1 {%k1}
; X86-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: vpaddd %ymm0, %ymm2, %ymm0
; X86-NEXT: vpaddd %ymm0, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_d_256:
; X64: # %bb.0:
; X64-NEXT: vpconflictd %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %ymm0, %ymm1 {%k1}
; X64-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: vpaddd %ymm0, %ymm2, %ymm0
; X64-NEXT: vpaddd %ymm0, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%1 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x1
%4 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %x0)
%5 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %x0)
%6 = bitcast i8 %x2 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
ret <8 x i32> %3
}
define <8 x i32> @test_int_x86_avx512_maskz_vpconflict_d_256(<8 x i32> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_d_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_d_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictd %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: retq
%1 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> zeroinitializer
ret <8 x i32> %3
}
define <2 x i64> @test_int_x86_avx512_vpconflict_q_128(<2 x i64> %x0, <2 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_q_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictq %xmm0, %xmm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %x0)
ret <2 x i64> %1
}
define <2 x i64> @test_int_x86_avx512_mask_vpconflict_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_q_128:
; X86: # %bb.0:
; X86-NEXT: vpconflictq %xmm0, %xmm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %xmm0, %xmm1 {%k1}
; X86-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0
; X86-NEXT: vpaddq %xmm0, %xmm1, %xmm0
; X86-NEXT: vmovdqa %xmm1, %xmm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_q_128:
; X64: # %bb.0:
; X64-NEXT: vpconflictq %xmm0, %xmm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %xmm0, %xmm1 {%k1}
; X64-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0
; X64-NEXT: vpaddq %xmm0, %xmm1, %xmm0
; X64-NEXT: vmovdqa %xmm1, %xmm0
; X64-NEXT: retq
%1 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract1, <2 x i64> %1, <2 x i64> %x1
%4 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %x0)
%5 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %x0)
%6 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <2 x i32> <i32 0, i32 1>
%7 = select <2 x i1> %extract, <2 x i64> %5, <2 x i64> zeroinitializer
%res3 = add <2 x i64> %3, %4
%res4 = add <2 x i64> %7, %res3
ret <2 x i64> %res4
ret <2 x i64> %3
}
define <2 x i64> @test_int_x86_avx512_maskz_vpconflict_q_128(<2 x i64> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_q_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_q_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %xmm0, %xmm0 {%k1} {z}
; X64-NEXT: retq
%1 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract1, <2 x i64> %1, <2 x i64> zeroinitializer
ret <2 x i64> %3
}
define <4 x i64> @test_int_x86_avx512_vpconflict_q_256(<4 x i64> %x0, <4 x i64> %x1) {
; CHECK-LABEL: test_int_x86_avx512_vpconflict_q_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpconflictq %ymm0, %ymm0
; CHECK-NEXT: ret{{[l|q]}}
%1 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %x0)
ret <4 x i64> %1
}
define <4 x i64> @test_int_x86_avx512_mask_vpconflict_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_mask_vpconflict_q_256:
; X86: # %bb.0:
; X86-NEXT: vpconflictq %ymm0, %ymm2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %ymm0, %ymm1 {%k1}
; X86-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0
; X86-NEXT: vpaddq %ymm0, %ymm1, %ymm0
; X86-NEXT: vmovdqa %ymm1, %ymm0
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_mask_vpconflict_q_256:
; X64: # %bb.0:
; X64-NEXT: vpconflictq %ymm0, %ymm2
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %ymm0, %ymm1 {%k1}
; X64-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0
; X64-NEXT: vpaddq %ymm0, %ymm1, %ymm0
; X64-NEXT: vmovdqa %ymm1, %ymm0
; X64-NEXT: retq
%1 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i64> %1, <4 x i64> %x1
%4 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %x0)
%5 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %x0)
%6 = bitcast i8 %x2 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract, <4 x i64> %5, <4 x i64> zeroinitializer
%res3 = add <4 x i64> %3, %4
%res4 = add <4 x i64> %7, %res3
ret <4 x i64> %res4
ret <4 x i64> %3
}
define <4 x i64> @test_int_x86_avx512_maskz_vpconflict_q_256(<4 x i64> %x0, i8 %x2) {
; X86-LABEL: test_int_x86_avx512_maskz_vpconflict_q_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax
; X86-NEXT: kmovw %eax, %k1
; X86-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X86-NEXT: retl
;
; X64-LABEL: test_int_x86_avx512_maskz_vpconflict_q_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1
; X64-NEXT: vpconflictq %ymm0, %ymm0 {%k1} {z}
; X64-NEXT: retq
%1 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %x0)
%2 = bitcast i8 %x2 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i64> %1, <4 x i64> zeroinitializer
ret <4 x i64> %3
}
declare <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32>)

6
llvm/test/CodeGen/X86/avx512dqvl-intrinsics.ll
View File

@ -1381,7 +1381,7 @@ define <8 x float>@test_int_x86_avx512_mask_reduce_ps_256(<8 x float> %x0, <8 x
; X86: # %bb.0:
; X86-NEXT: kmovb {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vreduceps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x56,0xc8,0x0b]
; X86-NEXT: vreduceps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0b]
; X86-NEXT: vreduceps $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0c]
; X86-NEXT: vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1389,11 +1389,11 @@ define <8 x float>@test_int_x86_avx512_mask_reduce_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vreduceps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x56,0xc8,0x0b]
; X64-NEXT: vreduceps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0b]
; X64-NEXT: vreduceps $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0c]
; X64-NEXT: vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
%res1 = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 12, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
ret <8 x float> %res2
}

124
llvm/test/CodeGen/X86/avx512ifma-intrinsics-upgrade.ll
View File

@ -4,46 +4,32 @@
declare <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
define <8 x i64>@test_int_x86_avx512_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
ret <8 x i64> %res
}
define <8 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res3 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
%res4 = add <8 x i64> %res, %res1
%res5 = add <8 x i64> %res3, %res2
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %res
}
declare <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
@ -51,87 +37,49 @@ declare <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64>, <8 x i64>,
define <8 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res2 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res3 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
%res4 = add <8 x i64> %res, %res1
%res5 = add <8 x i64> %res3, %res2
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %res
}
declare <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
define <8 x i64>@test_int_x86_avx512_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
ret <8 x i64> %res
}
define <8 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res2 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res3 = call <8 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
%res4 = add <8 x i64> %res, %res1
%res5 = add <8 x i64> %res3, %res2
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %res
}
declare <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
@ -139,43 +87,19 @@ declare <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64>, <8 x i64>,
define <8 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res2 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer, i8 %x3)
%res3 = call <8 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
%res4 = add <8 x i64> %res, %res1
%res5 = add <8 x i64> %res3, %res2
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %res
}
define <8 x i64>@test_int_x86_avx512_vpmadd52h_uq_512_load(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2ptr) {

140
llvm/test/CodeGen/X86/avx512ifma-intrinsics.ll
View File

@ -4,198 +4,106 @@
declare <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64>, <8 x i64>, <8 x i64>)
define <8 x i64>@test_int_x86_avx512_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
ret <8 x i64> %1
}
define <8 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb5,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x0
%4 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> %x0
%7 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%9 = select <8 x i1> %8, <8 x i64> %7, <8 x i64> zeroinitializer
%10 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%res4 = add <8 x i64> %3, %6
%res5 = add <8 x i64> %10, %9
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %3
}
define <8 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb5,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> zeroinitializer
%4 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> zeroinitializer
%7 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%9 = select <8 x i1> %8, <8 x i64> %7, <8 x i64> zeroinitializer
%10 = call <8 x i64> @llvm.x86.avx512.vpmadd52h.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%res4 = add <8 x i64> %3, %6
%res5 = add <8 x i64> %10, %9
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %3
}
declare <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64>, <8 x i64>, <8 x i64>)
define <8 x i64>@test_int_x86_avx512_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
ret <8 x i64> %1
}
define <8 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0xb4,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x0
%4 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> %x0
%7 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%9 = select <8 x i1> %8, <8 x i64> %7, <8 x i64> zeroinitializer
%10 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%res4 = add <8 x i64> %3, %6
%res5 = add <8 x i64> %10, %9
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %3
}
define <8 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_512:
; X86: # %bb.0:
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_512:
; X64: # %bb.0:
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm3 # encoding: [0x62,0xf2,0xf5,0x48,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %zmm2, %zmm1, %zmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0xb4,0xd2]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> zeroinitializer
%4 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> zeroinitializer
%7 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> zeroinitializer, <8 x i64> %x1, <8 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%9 = select <8 x i1> %8, <8 x i64> %7, <8 x i64> zeroinitializer
%10 = call <8 x i64> @llvm.x86.avx512.vpmadd52l.uq.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2)
%res4 = add <8 x i64> %3, %6
%res5 = add <8 x i64> %10, %9
%res6 = add <8 x i64> %res5, %res4
ret <8 x i64> %res6
ret <8 x i64> %3
}
define <8 x i64>@test_int_x86_avx512_vpmadd52h_uq_512_load(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2ptr) {

248
llvm/test/CodeGen/X86/avx512ifmavl-intrinsics-upgrade.ll
View File

@ -4,90 +4,62 @@
declare <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
define <2 x i64>@test_int_x86_avx512_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
ret <2 x i64> %res
}
define <2 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res2 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res3 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
%res4 = add <2 x i64> %res, %res1
%res5 = add <2 x i64> %res3, %res2
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
define <4 x i64>@test_int_x86_avx512_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
ret <4 x i64> %res
}
define <4 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res3 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
%res4 = add <4 x i64> %res, %res1
%res5 = add <4 x i64> %res3, %res2
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %res
}
declare <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
@ -95,43 +67,19 @@ declare <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64>, <2 x i64>,
define <2 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
%res1 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res2 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res3 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
%res4 = add <2 x i64> %res, %res1
%res5 = add <2 x i64> %res3, %res2
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
@ -139,131 +87,79 @@ declare <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64>, <4 x i64>,
define <4 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
%res1 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res2 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res3 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
%res4 = add <4 x i64> %res, %res1
%res5 = add <4 x i64> %res3, %res2
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %res
}
declare <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
define <2 x i64>@test_int_x86_avx512_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
ret <2 x i64> %res
}
define <2 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res2 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res3 = call <2 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
%res4 = add <2 x i64> %res, %res1
%res5 = add <2 x i64> %res3, %res2
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
define <4 x i64>@test_int_x86_avx512_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
ret <4 x i64> %res
}
define <4 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res3 = call <4 x i64> @llvm.x86.avx512.mask.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
%res4 = add <4 x i64> %res, %res1
%res5 = add <4 x i64> %res3, %res2
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %res
}
declare <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
@ -271,43 +167,19 @@ declare <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64>, <2 x i64>,
define <2 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
%res1 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res2 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
%res3 = call <2 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
%res4 = add <2 x i64> %res, %res1
%res5 = add <2 x i64> %res3, %res2
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %res
}
declare <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
@ -315,41 +187,17 @@ declare <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64>, <4 x i64>,
define <4 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
%res1 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res2 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
%res3 = call <4 x i64> @llvm.x86.avx512.maskz.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
%res4 = add <4 x i64> %res, %res1
%res5 = add <4 x i64> %res3, %res2
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %res
}

296
llvm/test/CodeGen/X86/avx512ifmavl-intrinsics.ll
View File

@ -4,416 +4,216 @@
declare <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64>, <2 x i64>, <2 x i64>)
define <2 x i64>@test_int_x86_avx512_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
ret <2 x i64> %1
}
define <2 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb5,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract2, <2 x i64> %1, <2 x i64> %x0
%4 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> %x0
%7 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <2 x i32> <i32 0, i32 1>
%9 = select <2 x i1> %extract, <2 x i64> %7, <2 x i64> zeroinitializer
%10 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%res4 = add <2 x i64> %3, %6
%res5 = add <2 x i64> %10, %9
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %3
}
declare <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64>, <4 x i64>, <4 x i64>)
define <4 x i64>@test_int_x86_avx512_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52h_uq_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
ret <4 x i64> %1
}
define <4 x i64>@test_int_x86_avx512_mask_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52h_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb5,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract2, <4 x i64> %1, <4 x i64> %x0
%4 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> %x0
%7 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = select <4 x i1> %extract, <4 x i64> %7, <4 x i64> zeroinitializer
%10 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%res4 = add <4 x i64> %3, %6
%res5 = add <4 x i64> %10, %9
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %3
}
define <2 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb5,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract2, <2 x i64> %1, <2 x i64> zeroinitializer
%4 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> zeroinitializer
%7 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <2 x i32> <i32 0, i32 1>
%9 = select <2 x i1> %extract, <2 x i64> %7, <2 x i64> zeroinitializer
%10 = call <2 x i64> @llvm.x86.avx512.vpmadd52h.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%res4 = add <2 x i64> %3, %6
%res5 = add <2 x i64> %10, %9
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %3
}
define <4 x i64>@test_int_x86_avx512_maskz_vpmadd52h_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52h_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb5,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52huq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb5,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract2, <4 x i64> %1, <4 x i64> zeroinitializer
%4 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> zeroinitializer
%7 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = select <4 x i1> %extract, <4 x i64> %7, <4 x i64> zeroinitializer
%10 = call <4 x i64> @llvm.x86.avx512.vpmadd52h.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%res4 = add <4 x i64> %3, %6
%res5 = add <4 x i64> %10, %9
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %3
}
declare <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64>, <2 x i64>, <2 x i64>)
define <2 x i64>@test_int_x86_avx512_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
ret <2 x i64> %1
}
define <2 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0xb4,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract2, <2 x i64> %1, <2 x i64> %x0
%4 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> %x0
%7 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <2 x i32> <i32 0, i32 1>
%9 = select <2 x i1> %extract, <2 x i64> %7, <2 x i64> zeroinitializer
%10 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%res4 = add <2 x i64> %3, %6
%res5 = add <2 x i64> %10, %9
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %3
}
declare <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64>, <4 x i64>, <4 x i64>)
define <4 x i64>@test_int_x86_avx512_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpmadd52l_uq_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
ret <4 x i64> %1
}
define <4 x i64>@test_int_x86_avx512_mask_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpmadd52l_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0xb4,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract2, <4 x i64> %1, <4 x i64> %x0
%4 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> %x0
%7 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = select <4 x i1> %extract, <4 x i64> %7, <4 x i64> zeroinitializer
%10 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%res4 = add <4 x i64> %3, %6
%res5 = add <4 x i64> %10, %9
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %3
}
define <2 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf2,0xf5,0x08,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %xmm2, %xmm1, %xmm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0xb4,0xd2]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract2, <2 x i64> %1, <2 x i64> zeroinitializer
%4 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> zeroinitializer
%7 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> zeroinitializer, <2 x i64> %x1, <2 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <2 x i32> <i32 0, i32 1>
%9 = select <2 x i1> %extract, <2 x i64> %7, <2 x i64> zeroinitializer
%10 = call <2 x i64> @llvm.x86.avx512.vpmadd52l.uq.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
%res4 = add <2 x i64> %3, %6
%res5 = add <2 x i64> %10, %9
%res6 = add <2 x i64> %res5, %res4
ret <2 x i64> %res6
ret <2 x i64> %3
}
define <4 x i64>@test_int_x86_avx512_maskz_vpmadd52l_uq_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xe2]
; X86-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_vpmadd52l_uq_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf2,0xf5,0x28,0xb4,0xda]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xe2]
; X64-NEXT: vpxor %xmm2, %xmm2, %xmm2 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xd2]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpmadd52luq %ymm2, %ymm1, %ymm2 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0xb4,0xd2]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%2 = bitcast i8 %x3 to <8 x i1>
%extract2 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract2, <4 x i64> %1, <4 x i64> zeroinitializer
%4 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer)
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> zeroinitializer
%7 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> zeroinitializer, <4 x i64> %x1, <4 x i64> zeroinitializer)
%8 = bitcast i8 %x3 to <8 x i1>
%extract = shufflevector <8 x i1> %8, <8 x i1> %8, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = select <4 x i1> %extract, <4 x i64> %7, <4 x i64> zeroinitializer
%10 = call <4 x i64> @llvm.x86.avx512.vpmadd52l.uq.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
%res4 = add <4 x i64> %3, %6
%res5 = add <4 x i64> %10, %9
%res6 = add <4 x i64> %res5, %res4
ret <4 x i64> %res6
ret <4 x i64> %3
}

265
llvm/test/CodeGen/X86/avx512vbmi2-intrinsics-upgrade.ll
View File

@ -367,21 +367,21 @@ declare <64 x i8> @llvm.x86.avx512.mask.compress.b.512(<64 x i8> %data, <64 x i8
define <16 x i32>@test_int_x86_avx512_mask_vpshld_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_d_512:
; X86: # %bb.0:
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_d_512:
; X64: # %bb.0:
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22, <16 x i32> %x3, i16 %x4)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22, <16 x i32> %x3, i16 -1)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 23, <16 x i32> %x3, i16 -1)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
@ -390,22 +390,22 @@ declare <16 x i32> @llvm.x86.avx512.mask.vpshld.d.512(<16 x i32>, <16 x i32>, i3
define <8 x i64>@test_int_x86_avx512_mask_vpshld_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_q_512:
; X86: # %bb.0:
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_q_512:
; X64: # %bb.0:
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22, <8 x i64> %x3, i8 %x4)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22, <8 x i64> %x3, i8 -1)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 23, <8 x i64> %x3, i8 -1)
%res2 = add <8 x i64> %res, %res1
ret <8 x i64> %res2
}
@ -414,21 +414,21 @@ declare <8 x i64> @llvm.x86.avx512.mask.vpshld.q.512(<8 x i64>, <8 x i64>, i32,
define <32 x i16>@test_int_x86_avx512_mask_vpshld_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_w_512:
; X86: # %bb.0:
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_w_512:
; X64: # %bb.0:
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6, <32 x i16> %x3, i32 %x4)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6, <32 x i16> %x3, i32 -1)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 7, <32 x i16> %x3, i32 -1)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res2
}
@ -437,21 +437,21 @@ declare <32 x i16> @llvm.x86.avx512.mask.vpshld.w.512(<32 x i16>, <32 x i16>, i3
define <16 x i32>@test_int_x86_avx512_mask_vpshrd_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_d_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_d_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22, <16 x i32> %x3, i16 %x4)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22, <16 x i32> %x3, i16 -1)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 23, <16 x i32> %x3, i16 -1)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
@ -460,22 +460,22 @@ declare <16 x i32> @llvm.x86.avx512.mask.vpshrd.d.512(<16 x i32>, <16 x i32>, i3
define <8 x i64>@test_int_x86_avx512_mask_vpshrd_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_q_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_q_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22, <8 x i64> %x3, i8 %x4)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22, <8 x i64> %x3, i8 -1)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 23, <8 x i64> %x3, i8 -1)
%res2 = add <8 x i64> %res, %res1
ret <8 x i64> %res2
}
@ -484,21 +484,21 @@ declare <8 x i64> @llvm.x86.avx512.mask.vpshrd.q.512(<8 x i64>, <8 x i64>, i32,
define <32 x i16>@test_int_x86_avx512_mask_vpshrd_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_w_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_w_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6, <32 x i16> %x3, i32 %x4)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6, <32 x i16> %x3, i32 -1)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 7, <32 x i16> %x3, i32 -1)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res2
}
@ -507,23 +507,23 @@ declare <32 x i16> @llvm.x86.avx512.mask.vpshrd.w.512(<32 x i16>, <32 x i16>, i3
define <16 x i32>@test_int_x86_avx512_mask_vpshld_d_512_2(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_d_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_d_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x3
%4 = call <16 x i32> @llvm.x86.avx512.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22)
%4 = call <16 x i32> @llvm.x86.avx512.vpshld.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 23)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
}
@ -532,24 +532,24 @@ declare <16 x i32> @llvm.x86.avx512.vpshld.d.512(<16 x i32>, <16 x i32>, i32)
define <8 x i64>@test_int_x86_avx512_mask_vpshld_q_512_2(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_q_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_q_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x3
%4 = call <8 x i64> @llvm.x86.avx512.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22)
%4 = call <8 x i64> @llvm.x86.avx512.vpshld.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 23)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
}
@ -558,23 +558,23 @@ declare <8 x i64> @llvm.x86.avx512.vpshld.q.512(<8 x i64>, <8 x i64>, i32)
define <32 x i16>@test_int_x86_avx512_mask_vpshld_w_512_2(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_w_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_w_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6)
%2 = bitcast i32 %x4 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x3
%4 = call <32 x i16> @llvm.x86.avx512.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6)
%4 = call <32 x i16> @llvm.x86.avx512.vpshld.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 7)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
}
@ -583,23 +583,23 @@ declare <32 x i16> @llvm.x86.avx512.vpshld.w.512(<32 x i16>, <32 x i16>, i32)
define <16 x i32>@test_int_x86_avx512_mask_vpshrd_d_512_2(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_d_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_d_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x3
%4 = call <16 x i32> @llvm.x86.avx512.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 22)
%4 = call <16 x i32> @llvm.x86.avx512.vpshrd.d.512(<16 x i32> %x0, <16 x i32> %x1, i32 23)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
}
@ -608,24 +608,24 @@ declare <16 x i32> @llvm.x86.avx512.vpshrd.d.512(<16 x i32>, <16 x i32>, i32)
define <8 x i64>@test_int_x86_avx512_mask_vpshrd_q_512_2(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_q_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_q_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.x86.avx512.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x3
%4 = call <8 x i64> @llvm.x86.avx512.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 22)
%4 = call <8 x i64> @llvm.x86.avx512.vpshrd.q.512(<8 x i64> %x0, <8 x i64> %x1, i32 23)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
}
@ -634,23 +634,23 @@ declare <8 x i64> @llvm.x86.avx512.vpshrd.q.512(<8 x i64>, <8 x i64>, i32)
define <32 x i16>@test_int_x86_avx512_mask_vpshrd_w_512_2(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_w_512_2:
; X86: # %bb.0:
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_w_512_2:
; X64: # %bb.0:
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.x86.avx512.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6)
%2 = bitcast i32 %x4 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x3
%4 = call <32 x i16> @llvm.x86.avx512.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 6)
%4 = call <32 x i16> @llvm.x86.avx512.vpshrd.w.512(<32 x i16> %x0, <32 x i16> %x1, i32 7)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
}
@ -659,6 +659,15 @@ declare <32 x i16> @llvm.x86.avx512.vpshrd.w.512(<32 x i16>, <32 x i16>, i32)
declare <16 x i32> @llvm.x86.avx512.mask.vpshrdv.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpshrdv.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpshrdv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshrdv_d_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshrdvd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x73,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshrdv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrdv_d_512:
; X86: # %bb.0:
@ -666,10 +675,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x73,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x73,0xe2]
; X86-NEXT: vpshrdvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x73,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -678,24 +684,28 @@ define <16 x i32>@test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x73,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x73,0xe2]
; X64-NEXT: vpshrdvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x73,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshrdv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshrdv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpshrdv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpshrdv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
declare <8 x i64> @llvm.x86.avx512.mask.vpshrdv.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
declare <8 x i64> @llvm.x86.avx512.maskz.vpshrdv.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
define <8 x i64>@test_int_x86_avx512_vpshrdv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshrdv_q_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshrdvq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0x73,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshrdv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
ret <8 x i64> %res
}
define <8 x i64>@test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2p, <8 x i64> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrdv_q_512:
; X86: # %bb.0:
@ -704,10 +714,7 @@ define <8 x i64>@test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64>
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvq (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x73,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x73,0xe2]
; X86-NEXT: vpshrdvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x73,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -716,61 +723,59 @@ define <8 x i64>@test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64>
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvq (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x73,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x73,0xe2]
; X64-NEXT: vpshrdvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x73,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i64>, <8 x i64>* %x2p
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshrdv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshrdv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 -1)
%res2 = call <8 x i64> @llvm.x86.avx512.maskz.vpshrdv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 %x3)
%res3 = add <8 x i64> %res, %res1
%res4 = add <8 x i64> %res2, %res3
ret <8 x i64> %res4
%res1 = call <8 x i64> @llvm.x86.avx512.maskz.vpshrdv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 %x3)
%res2 = add <8 x i64> %res, %res1
ret <8 x i64> %res2
}
declare <32 x i16> @llvm.x86.avx512.mask.vpshrdv.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
declare <32 x i16> @llvm.x86.avx512.maskz.vpshrdv.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_vpshrdv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshrdv_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshrdvw %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0x72,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshrdv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_vpshrdv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16>* %x2p, <32 x i16> %x4, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrdv_w_512:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvw (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x72,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x72,0xe2]
; X86-NEXT: vpshrdvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x72,0xc2]
; X86-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrdv_w_512:
; X64: # %bb.0:
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvw (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x72,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x72,0xe2]
; X64-NEXT: vpshrdvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x72,0xc2]
; X64-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <32 x i16>, <32 x i16>* %x2p
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshrdv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshrdv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 -1)
%res2 = call <32 x i16> @llvm.x86.avx512.maskz.vpshrdv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 %x3)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res2, %res3
ret <32 x i16> %res4
%res1 = call <32 x i16> @llvm.x86.avx512.maskz.vpshrdv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 %x3)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res1
}
declare <16 x i32> @llvm.x86.avx512.mask.vpshldv.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpshldv.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpshldv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshldv_d_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshldvd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x71,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshldv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshldv_d_512:
; X86: # %bb.0:
@ -778,10 +783,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x71,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x71,0xe2]
; X86-NEXT: vpshldvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x71,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -790,24 +792,28 @@ define <16 x i32>@test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x71,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x71,0xe2]
; X64-NEXT: vpshldvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x71,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpshldv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpshldv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpshldv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpshldv.d.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
declare <8 x i64> @llvm.x86.avx512.mask.vpshldv.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
declare <8 x i64> @llvm.x86.avx512.maskz.vpshldv.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
define <8 x i64>@test_int_x86_avx512_vpshldv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshldv_q_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshldvq %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0x71,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshldv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 -1)
ret <8 x i64> %res
}
define <8 x i64>@test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2p, <8 x i64> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshldv_q_512:
; X86: # %bb.0:
@ -816,10 +822,7 @@ define <8 x i64>@test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64>
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvq (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x71,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x71,0xe2]
; X86-NEXT: vpshldvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x71,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -828,24 +831,28 @@ define <8 x i64>@test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64>
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvq (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x71,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x71,0xe2]
; X64-NEXT: vpshldvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x71,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i64>, <8 x i64>* %x2p
%res = call <8 x i64> @llvm.x86.avx512.mask.vpshldv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x2, i8 %x3)
%res1 = call <8 x i64> @llvm.x86.avx512.mask.vpshldv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 -1)
%res2 = call <8 x i64> @llvm.x86.avx512.maskz.vpshldv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 %x3)
%res3 = add <8 x i64> %res, %res1
%res4 = add <8 x i64> %res2, %res3
ret <8 x i64> %res4
%res1 = call <8 x i64> @llvm.x86.avx512.maskz.vpshldv.q.512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4, i8 %x3)
%res2 = add <8 x i64> %res, %res1
ret <8 x i64> %res2
}
declare <32 x i16> @llvm.x86.avx512.mask.vpshldv.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
declare <32 x i16> @llvm.x86.avx512.maskz.vpshldv.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16>@test_int_x86_avx512_vpshldv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpshldv_w_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpshldvw %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0x70,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshldv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 -1)
ret <32 x i16> %res
}
define <32 x i16>@test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16>* %x2p, <32 x i16> %x4, i32 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpshldv_w_512:
; X86: # %bb.0:
@ -853,10 +860,7 @@ define <32 x i16>@test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x i
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvw (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x70,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x70,0xe2]
; X86-NEXT: vpshldvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x70,0xc2]
; X86-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -865,17 +869,12 @@ define <32 x i16>@test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x i
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvw (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x70,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x70,0xe2]
; X64-NEXT: vpshldvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x70,0xc2]
; X64-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <32 x i16>, <32 x i16>* %x2p
%res = call <32 x i16> @llvm.x86.avx512.mask.vpshldv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x2, i32 %x3)
%res1 = call <32 x i16> @llvm.x86.avx512.mask.vpshldv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 -1)
%res2 = call <32 x i16> @llvm.x86.avx512.maskz.vpshldv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 %x3)
%res3 = add <32 x i16> %res, %res1
%res4 = add <32 x i16> %res2, %res3
ret <32 x i16> %res4
%res1 = call <32 x i16> @llvm.x86.avx512.maskz.vpshldv.w.512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4, i32 %x3)
%res2 = add <32 x i16> %res, %res1
ret <32 x i16> %res2
}

156
llvm/test/CodeGen/X86/avx512vbmi2-intrinsics.ll
View File

@ -370,23 +370,23 @@ define <64 x i8> @test_compress_b_512(<64 x i8> %data) {
define <16 x i32> @test_int_x86_avx512_mask_vpshld_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_d_512:
; X86: # %bb.0:
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_d_512:
; X64: # %bb.0:
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshldd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x3
%4 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%4 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
}
@ -394,24 +394,24 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshld_d_512(<16 x i32> %x0, <16 x i
define <8 x i64> @test_int_x86_avx512_mask_vpshld_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_q_512:
; X86: # %bb.0:
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_q_512:
; X64: # %bb.0:
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x71,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshldq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x71,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> <i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x3
%4 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> <i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22>)
%4 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> <i64 23, i64 23, i64 23, i64 23, i64 23, i64 23, i64 23, i64 23>)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
}
@ -419,23 +419,23 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshld_q_512(<8 x i64> %x0, <8 x i64>
define <32 x i16> @test_int_x86_avx512_mask_vpshld_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_w_512:
; X86: # %bb.0:
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_w_512:
; X64: # %bb.0:
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x70,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshldw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x70,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i32 %x4 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x3
%4 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
}
@ -443,23 +443,23 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshld_w_512(<32 x i16> %x0, <32 x i
define <16 x i32> @test_int_x86_avx512_mask_vpshrd_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_d_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X86-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_d_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdd $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddd %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc3]
; X64-NEXT: vpshrdd $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0x7d,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddd %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %x1, <16 x i32> %x0, <16 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x3
%4 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %x1, <16 x i32> %x0, <16 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%4 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %x1, <16 x i32> %x0, <16 x i32> <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>)
%res2 = add <16 x i32> %3, %4
ret <16 x i32> %res2
}
@ -467,24 +467,24 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshrd_d_512(<16 x i32> %x0, <16 x i
define <8 x i64> @test_int_x86_avx512_mask_vpshrd_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_q_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X86-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X86-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X86-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_q_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdq $22, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x73,0xd1,0x16]
; X64-NEXT: vpaddq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc3]
; X64-NEXT: vpshrdq $23, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x73,0xc1,0x17]
; X64-NEXT: vpaddq %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %x1, <8 x i64> %x0, <8 x i64> <i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x3
%4 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %x1, <8 x i64> %x0, <8 x i64> <i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22, i64 22>)
%4 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %x1, <8 x i64> %x0, <8 x i64> <i64 23, i64 23, i64 23, i64 23, i64 23, i64 23, i64 23, i64 23>)
%res2 = add <8 x i64> %3, %4
ret <8 x i64> %res2
}
@ -492,23 +492,23 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshrd_q_512(<8 x i64> %x0, <8 x i64>
define <32 x i16> @test_int_x86_avx512_mask_vpshrd_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x3, i32 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_w_512:
; X86: # %bb.0:
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X86-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X86-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X86-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_w_512:
; X64: # %bb.0:
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdw $6, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0x72,0xd1,0x06]
; X64-NEXT: vpaddw %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc3]
; X64-NEXT: vpshrdw $7, %zmm1, %zmm0, %zmm0 # encoding: [0x62,0xf3,0xfd,0x48,0x72,0xc1,0x07]
; X64-NEXT: vpaddw %zmm0, %zmm2, %zmm0 # encoding: [0x62,0xf1,0x6d,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %x1, <32 x i16> %x0, <32 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i32 %x4 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x3
%4 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %x1, <32 x i16> %x0, <32 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %x1, <32 x i16> %x0, <32 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <32 x i16> %3, %4
ret <32 x i16> %res2
}
@ -520,10 +520,7 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x73,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x73,0xe2]
; X86-NEXT: vpshrdvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x73,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -532,10 +529,7 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x73,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x73,0xe2]
; X64-NEXT: vpshrdvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x73,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -543,12 +537,10 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshrdv_d_512(<16 x i32> %x0, <16 x
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %x1, <16 x i32> %x0, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %x1, <16 x i32> %x0, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}
define <8 x i64> @test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2p, <8 x i64> %x4, i8 %x3) {
@ -559,10 +551,7 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvq (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x73,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x73,0xe2]
; X86-NEXT: vpshrdvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x73,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -571,10 +560,7 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvq (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x73,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x73,0xe2]
; X64-NEXT: vpshrdvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x73,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i64>, <8 x i64>* %x2p
@ -582,12 +568,10 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshrdv_q_512(<8 x i64> %x0, <8 x i64
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x0
%4 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %x1, <8 x i64> %x0, <8 x i64> %x4)
%5 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %x1, <8 x i64> %x0, <8 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i64> %5, <8 x i64> zeroinitializer
%res3 = add <8 x i64> %3, %4
%res4 = add <8 x i64> %7, %res3
ret <8 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> zeroinitializer
%res3 = add <8 x i64> %3, %6
ret <8 x i64> %res3
}
define <32 x i16> @test_int_x86_avx512_mask_vpshrdv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16>* %x2p, <32 x i16> %x4, i32 %x3) {
@ -597,10 +581,7 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshrdv_w_512(<32 x i16> %x0, <32 x
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshrdvw (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x72,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshrdvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x72,0xe2]
; X86-NEXT: vpshrdvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x72,0xc2]
; X86-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -609,10 +590,7 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshrdv_w_512(<32 x i16> %x0, <32 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshrdvw (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x72,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshrdvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x72,0xe2]
; X64-NEXT: vpshrdvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x72,0xc2]
; X64-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <32 x i16>, <32 x i16>* %x2p
@ -620,12 +598,10 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshrdv_w_512(<32 x i16> %x0, <32 x
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x0
%4 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %x1, <32 x i16> %x0, <32 x i16> %x4)
%5 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %x1, <32 x i16> %x0, <32 x i16> %x4)
%6 = bitcast i32 %x3 to <32 x i1>
%7 = select <32 x i1> %6, <32 x i16> %5, <32 x i16> zeroinitializer
%res3 = add <32 x i16> %3, %4
%res4 = add <32 x i16> %7, %res3
ret <32 x i16> %res4
%5 = bitcast i32 %x3 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%res3 = add <32 x i16> %3, %6
ret <32 x i16> %res3
}
define <16 x i32> @test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
@ -635,10 +611,7 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x71,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x71,0xe2]
; X86-NEXT: vpshldvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x71,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -647,10 +620,7 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x71,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x71,0xe2]
; X64-NEXT: vpshldvd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x71,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -658,12 +628,10 @@ define <16 x i32> @test_int_x86_avx512_mask_vpshldv_d_512(<16 x i32> %x0, <16 x
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}
define <8 x i64> @test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64> %x1, <8 x i64>* %x2p, <8 x i64> %x4, i8 %x3) {
@ -674,10 +642,7 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvq (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x71,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x71,0xe2]
; X86-NEXT: vpshldvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x71,0xc2]
; X86-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X86-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -686,10 +651,7 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvq (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x71,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvq %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x71,0xe2]
; X64-NEXT: vpshldvq %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x71,0xc2]
; X64-NEXT: vpaddq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xd4,0xc0]
; X64-NEXT: vpaddq %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i64>, <8 x i64>* %x2p
@ -697,12 +659,10 @@ define <8 x i64> @test_int_x86_avx512_mask_vpshldv_q_512(<8 x i64> %x0, <8 x i64
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i64> %1, <8 x i64> %x0
%4 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4)
%5 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %x0, <8 x i64> %x1, <8 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i64> %5, <8 x i64> zeroinitializer
%res3 = add <8 x i64> %3, %4
%res4 = add <8 x i64> %7, %res3
ret <8 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i64> %4, <8 x i64> zeroinitializer
%res3 = add <8 x i64> %3, %6
ret <8 x i64> %res3
}
define <32 x i16> @test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x i16> %x1, <32 x i16>* %x2p, <32 x i16> %x4, i32 %x3) {
@ -712,10 +672,7 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x
; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpshldvw (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x70,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpshldvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x70,0xe2]
; X86-NEXT: vpshldvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x70,0xc2]
; X86-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X86-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -724,10 +681,7 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpshldvw (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x49,0x70,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpshldvw %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0xf5,0x48,0x70,0xe2]
; X64-NEXT: vpshldvw %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xc9,0x70,0xc2]
; X64-NEXT: vpaddw %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfd,0xc0]
; X64-NEXT: vpaddw %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <32 x i16>, <32 x i16>* %x2p
@ -735,12 +689,10 @@ define <32 x i16> @test_int_x86_avx512_mask_vpshldv_w_512(<32 x i16> %x0, <32 x
%2 = bitcast i32 %x3 to <32 x i1>
%3 = select <32 x i1> %2, <32 x i16> %1, <32 x i16> %x0
%4 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4)
%5 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %x0, <32 x i16> %x1, <32 x i16> %x4)
%6 = bitcast i32 %x3 to <32 x i1>
%7 = select <32 x i1> %6, <32 x i16> %5, <32 x i16> zeroinitializer
%res3 = add <32 x i16> %3, %4
%res4 = add <32 x i16> %7, %res3
ret <32 x i16> %res4
%5 = bitcast i32 %x3 to <32 x i1>
%6 = select <32 x i1> %5, <32 x i16> %4, <32 x i16> zeroinitializer
%res3 = add <32 x i16> %3, %6
ret <32 x i16> %res3
}
declare <16 x i32> @llvm.fshl.v16i32(<16 x i32>, <16 x i32>, <16 x i32>)

524
llvm/test/CodeGen/X86/avx512vbmi2vl-intrinsics-upgrade.ll
View File

File diff suppressed because it is too large Load Diff

332
llvm/test/CodeGen/X86/avx512vbmi2vl-intrinsics.ll
View File

@ -734,21 +734,21 @@ define <32 x i8> @test_compress_b_256(<32 x i8> %data) {
define <4 x i32> @test_int_x86_avx512_mask_vpshld_d_128(<4 x i32> %x0, <4 x i32> %x1,<4 x i32> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_d_128:
; X86: # %bb.0:
; X86-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x71,0xd1,0x16]
; X86-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x71,0xc1,0x16]
; X86-NEXT: vpshldd $23, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x71,0xd9,0x17]
; X86-NEXT: vpshldd $24, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x71,0xc1,0x18]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_d_128:
; X64: # %bb.0:
; X64-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x71,0xd1,0x16]
; X64-NEXT: vpshldd $22, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x71,0xc1,0x16]
; X64-NEXT: vpshldd $23, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x71,0xd9,0x17]
; X64-NEXT: vpshldd $24, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x71,0xc1,0x18]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
@ -756,8 +756,8 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshld_d_128(<4 x i32> %x0, <4 x i32>
%2 = bitcast i8 %x4 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> %x3
%4 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> <i32 22, i32 22, i32 22, i32 22>)
%5 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> <i32 22, i32 22, i32 22, i32 22>)
%4 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> <i32 23, i32 23, i32 23, i32 23>)
%5 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> <i32 24, i32 24, i32 24, i32 24>)
%6 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract, <4 x i32> %5, <4 x i32> zeroinitializer
@ -769,24 +769,24 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshld_d_128(<4 x i32> %x0, <4 x i32>
define <8 x i32> @test_int_x86_avx512_mask_vpshld_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_d_256:
; X86: # %bb.0:
; X86-NEXT: vpshldd $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldd $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x71,0xd1,0x16]
; X86-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X86-NEXT: vpshldd $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x71,0xc1,0x17]
; X86-NEXT: vpaddd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_d_256:
; X64: # %bb.0:
; X64-NEXT: vpshldd $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldd $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x71,0xd1,0x16]
; X64-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X64-NEXT: vpshldd $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x71,0xc1,0x17]
; X64-NEXT: vpaddd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x3
%4 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%4 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
}
@ -794,25 +794,25 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshld_d_256(<8 x i32> %x0, <8 x i32>
define <2 x i64> @test_int_x86_avx512_mask_vpshld_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_q_128:
; X86: # %bb.0:
; X86-NEXT: vpshldq $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldq $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x71,0xd1,0x16]
; X86-NEXT: vpaddq %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc3]
; X86-NEXT: vpshldq $23, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x71,0xc1,0x17]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_q_128:
; X64: # %bb.0:
; X64-NEXT: vpshldq $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldq $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x71,0xd1,0x16]
; X64-NEXT: vpaddq %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc3]
; X64-NEXT: vpshldq $23, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x71,0xc1,0x17]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> <i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x3
%4 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> <i64 22, i64 22>)
%4 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> <i64 23, i64 23>)
%res2 = add <2 x i64> %3, %4
ret <2 x i64> %res2
}
@ -820,25 +820,25 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshld_q_128(<2 x i64> %x0, <2 x i64>
define <4 x i64> @test_int_x86_avx512_mask_vpshld_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_q_256:
; X86: # %bb.0:
; X86-NEXT: vpshldq $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x71,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldq $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x71,0xd1,0x16]
; X86-NEXT: vpaddq %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc3]
; X86-NEXT: vpshldq $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x71,0xc1,0x17]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_q_256:
; X64: # %bb.0:
; X64-NEXT: vpshldq $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x71,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldq $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x71,0xd1,0x16]
; X64-NEXT: vpaddq %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc3]
; X64-NEXT: vpshldq $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x71,0xc1,0x17]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> <i64 22, i64 22, i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x3
%4 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> <i64 22, i64 22, i64 22, i64 22>)
%4 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> <i64 23, i64 23, i64 23, i64 23>)
%res2 = add <4 x i64> %3, %4
ret <4 x i64> %res2
}
@ -846,24 +846,24 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshld_q_256(<4 x i64> %x0, <4 x i64>
define <8 x i16> @test_int_x86_avx512_mask_vpshld_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_w_128:
; X86: # %bb.0:
; X86-NEXT: vpshldw $6, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x70,0xd9,0x06]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshldw $6, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x70,0xd1,0x06]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vpshldw $7, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x70,0xc1,0x07]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_w_128:
; X64: # %bb.0:
; X64-NEXT: vpshldw $6, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x70,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldw $6, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x70,0xd1,0x06]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vpshldw $7, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x70,0xc1,0x07]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x3
%4 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
}
@ -871,23 +871,23 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshld_w_128(<8 x i16> %x0, <8 x i16>
define <16 x i16> @test_int_x86_avx512_mask_vpshld_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshld_w_256:
; X86: # %bb.0:
; X86-NEXT: vpshldw $6, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x70,0xd9,0x06]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshldw $6, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x70,0xd1,0x06]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vpshldw $7, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x70,0xc1,0x07]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshld_w_256:
; X64: # %bb.0:
; X64-NEXT: vpshldw $6, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x70,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshldw $6, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x70,0xd1,0x06]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vpshldw $7, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x70,0xc1,0x07]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x3
%4 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
}
@ -895,21 +895,21 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshld_w_256(<16 x i16> %x0, <16 x i
define <4 x i32> @test_int_x86_avx512_mask_vpshrd_d_128(<4 x i32> %x0, <4 x i32> %x1,<4 x i32> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_d_128:
; X86: # %bb.0:
; X86-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x73,0xd1,0x16]
; X86-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x73,0xc1,0x16]
; X86-NEXT: vpshrdd $23, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x73,0xd9,0x17]
; X86-NEXT: vpshrdd $24, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x73,0xc1,0x18]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_d_128:
; X64: # %bb.0:
; X64-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x73,0xd1,0x16]
; X64-NEXT: vpshrdd $22, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x73,0xc1,0x16]
; X64-NEXT: vpshrdd $23, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0x7d,0x08,0x73,0xd9,0x17]
; X64-NEXT: vpshrdd $24, %xmm1, %xmm0, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x7d,0x89,0x73,0xc1,0x18]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
@ -917,8 +917,8 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshrd_d_128(<4 x i32> %x0, <4 x i32>
%2 = bitcast i8 %x4 to <8 x i1>
%extract1 = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract1, <4 x i32> %1, <4 x i32> %x3
%4 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> <i32 22, i32 22, i32 22, i32 22>)
%5 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> <i32 22, i32 22, i32 22, i32 22>)
%4 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> <i32 23, i32 23, i32 23, i32 23>)
%5 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> <i32 24, i32 24, i32 24, i32 24>)
%6 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract, <4 x i32> %5, <4 x i32> zeroinitializer
@ -930,24 +930,24 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshrd_d_128(<4 x i32> %x0, <4 x i32>
define <8 x i32> @test_int_x86_avx512_mask_vpshrd_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_d_256:
; X86: # %bb.0:
; X86-NEXT: vpshrdd $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdd $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x73,0xd1,0x16]
; X86-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X86-NEXT: vpshrdd $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x73,0xc1,0x17]
; X86-NEXT: vpaddd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_d_256:
; X64: # %bb.0:
; X64-NEXT: vpshrdd $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0x7d,0x28,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdd $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x73,0xd1,0x16]
; X64-NEXT: vpaddd %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc3]
; X64-NEXT: vpshrdd $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x73,0xc1,0x17]
; X64-NEXT: vpaddd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x3
%4 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> <i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22, i32 22>)
%4 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> <i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23, i32 23>)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
}
@ -955,25 +955,25 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshrd_d_256(<8 x i32> %x0, <8 x i32>
define <2 x i64> @test_int_x86_avx512_mask_vpshrd_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_q_128:
; X86: # %bb.0:
; X86-NEXT: vpshrdq $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdq $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x73,0xd1,0x16]
; X86-NEXT: vpaddq %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc3]
; X86-NEXT: vpshrdq $23, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x73,0xc1,0x17]
; X86-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_q_128:
; X64: # %bb.0:
; X64-NEXT: vpshrdq $22, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdq $22, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x73,0xd1,0x16]
; X64-NEXT: vpaddq %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc3]
; X64-NEXT: vpshrdq $23, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x73,0xc1,0x17]
; X64-NEXT: vpaddq %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> <i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x3
%4 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> <i64 22, i64 22>)
%4 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> <i64 23, i64 23>)
%res2 = add <2 x i64> %3, %4
ret <2 x i64> %res2
}
@ -981,25 +981,25 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshrd_q_128(<2 x i64> %x0, <2 x i64>
define <4 x i64> @test_int_x86_avx512_mask_vpshrd_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_q_256:
; X86: # %bb.0:
; X86-NEXT: vpshrdq $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x73,0xd9,0x16]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdq $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x73,0xd1,0x16]
; X86-NEXT: vpaddq %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc3]
; X86-NEXT: vpshrdq $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x73,0xc1,0x17]
; X86-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_q_256:
; X64: # %bb.0:
; X64-NEXT: vpshrdq $22, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x73,0xd9,0x16]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdq $22, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x73,0xd1,0x16]
; X64-NEXT: vpaddq %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc3]
; X64-NEXT: vpshrdq $23, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x73,0xc1,0x17]
; X64-NEXT: vpaddq %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> <i64 22, i64 22, i64 22, i64 22>)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x3
%4 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> <i64 22, i64 22, i64 22, i64 22>)
%4 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> <i64 23, i64 23, i64 23, i64 23>)
%res2 = add <4 x i64> %3, %4
ret <4 x i64> %res2
}
@ -1007,24 +1007,24 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshrd_q_256(<4 x i64> %x0, <4 x i64>
define <8 x i16> @test_int_x86_avx512_mask_vpshrd_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x3, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_w_128:
; X86: # %bb.0:
; X86-NEXT: vpshrdw $6, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x72,0xd9,0x06]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovd %eax, %k1 # encoding: [0xc5,0xfb,0x92,0xc8]
; X86-NEXT: vpshrdw $6, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x72,0xd1,0x06]
; X86-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X86-NEXT: vpshrdw $7, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x72,0xc1,0x07]
; X86-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_w_128:
; X64: # %bb.0:
; X64-NEXT: vpshrdw $6, %xmm1, %xmm0, %xmm3 # encoding: [0x62,0xf3,0xfd,0x08,0x72,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdw $6, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x72,0xd1,0x06]
; X64-NEXT: vpaddw %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc3]
; X64-NEXT: vpshrdw $7, %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x72,0xc1,0x07]
; X64-NEXT: vpaddw %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x3
%4 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <8 x i16> %3, %4
ret <8 x i16> %res2
}
@ -1032,23 +1032,23 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshrd_w_128(<8 x i16> %x0, <8 x i16>
define <16 x i16> @test_int_x86_avx512_mask_vpshrd_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x3, i16 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_vpshrd_w_256:
; X86: # %bb.0:
; X86-NEXT: vpshrdw $6, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x72,0xd9,0x06]
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
; X86-NEXT: vpshrdw $6, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x72,0xd1,0x06]
; X86-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X86-NEXT: vpshrdw $7, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x72,0xc1,0x07]
; X86-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_vpshrd_w_256:
; X64: # %bb.0:
; X64-NEXT: vpshrdw $6, %ymm1, %ymm0, %ymm3 # encoding: [0x62,0xf3,0xfd,0x28,0x72,0xd9,0x06]
; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
; X64-NEXT: vpshrdw $6, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x72,0xd1,0x06]
; X64-NEXT: vpaddw %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc3]
; X64-NEXT: vpshrdw $7, %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x72,0xc1,0x07]
; X64-NEXT: vpaddw %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%2 = bitcast i16 %x4 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x3
%4 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> <i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6, i16 6>)
%4 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>)
%res2 = add <16 x i16> %3, %4
ret <16 x i16> %res2
}
@ -1061,10 +1061,7 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshrdv_d_256(<8 x i32> %x0, <8 x i32
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshrdvd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x73,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshrdvd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x73,0xe2]
; X86-NEXT: vpshrdvd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x73,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1073,10 +1070,7 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshrdv_d_256(<8 x i32> %x0, <8 x i32
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshrdvd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x73,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshrdvd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x73,0xe2]
; X64-NEXT: vpshrdvd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x73,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -1084,12 +1078,10 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshrdv_d_256(<8 x i32> %x0, <8 x i32
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
define <4 x i32> @test_int_x86_avx512_mask_vpshrdv_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
@ -1100,10 +1092,7 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshrdv_d_128(<4 x i32> %x0, <4 x i32
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshrdvd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x73,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshrdvd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x73,0xe2]
; X86-NEXT: vpshrdvd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x73,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1112,10 +1101,7 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshrdv_d_128(<4 x i32> %x0, <4 x i32
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshrdvd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x73,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshrdvd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x73,0xe2]
; X64-NEXT: vpshrdvd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x73,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -1124,13 +1110,11 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshrdv_d_128(<4 x i32> %x0, <4 x i32
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}
define <4 x i64> @test_int_x86_avx512_mask_vpshrdv_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64>* %x2p, <4 x i64> %x4, i8 %x3) {
@ -1141,10 +1125,7 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshrdv_q_256(<4 x i64> %x0, <4 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshrdvq (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x73,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshrdvq %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x73,0xe2]
; X86-NEXT: vpshrdvq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x73,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1153,10 +1134,7 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshrdv_q_256(<4 x i64> %x0, <4 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshrdvq (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x73,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshrdvq %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x73,0xe2]
; X64-NEXT: vpshrdvq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x73,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i64>, <4 x i64>* %x2p
@ -1165,13 +1143,11 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshrdv_q_256(<4 x i64> %x0, <4 x i64
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x0
%4 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> %x4)
%5 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i64> %5, <4 x i64> zeroinitializer
%res3 = add <4 x i64> %3, %4
%res4 = add <4 x i64> %7, %res3
ret <4 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> zeroinitializer
%res3 = add <4 x i64> %3, %6
ret <4 x i64> %res3
}
define <2 x i64> @test_int_x86_avx512_mask_vpshrdv_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64>* %x2p, <2 x i64> %x4, i8 %x3) {
@ -1182,10 +1158,7 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshrdv_q_128(<2 x i64> %x0, <2 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshrdvq (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x73,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshrdvq %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x73,0xe2]
; X86-NEXT: vpshrdvq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x73,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1194,10 +1167,7 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshrdv_q_128(<2 x i64> %x0, <2 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshrdvq (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x73,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshrdvq %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x73,0xe2]
; X64-NEXT: vpshrdvq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x73,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <2 x i64>, <2 x i64>* %x2p
@ -1206,13 +1176,11 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshrdv_q_128(<2 x i64> %x0, <2 x i64
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x0
%4 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> %x4)
%5 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <2 x i32> <i32 0, i32 1>
%7 = select <2 x i1> %extract1, <2 x i64> %5, <2 x i64> zeroinitializer
%res3 = add <2 x i64> %3, %4
%res4 = add <2 x i64> %7, %res3
ret <2 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> zeroinitializer
%res3 = add <2 x i64> %3, %6
ret <2 x i64> %res3
}
define <16 x i16> @test_int_x86_avx512_mask_vpshrdv_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16>* %x2p, <16 x i16> %x4, i16 %x3) {
@ -1222,10 +1190,7 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshrdv_w_256(<16 x i16> %x0, <16 x
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshrdvw (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x72,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshrdvw %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x72,0xe2]
; X86-NEXT: vpshrdvw %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x72,0xc2]
; X86-NEXT: vpaddw %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfd,0xc0]
; X86-NEXT: vpshrdvw %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0x72,0xc2]
; X86-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1234,10 +1199,7 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshrdv_w_256(<16 x i16> %x0, <16 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshrdvw (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x72,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshrdvw %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x72,0xe2]
; X64-NEXT: vpshrdvw %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x72,0xc2]
; X64-NEXT: vpaddw %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfd,0xc0]
; X64-NEXT: vpshrdvw %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0xf5,0x28,0x72,0xc2]
; X64-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i16>, <16 x i16>* %x2p
@ -1245,12 +1207,10 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshrdv_w_256(<16 x i16> %x0, <16 x
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x0
%4 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> %x4)
%5 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i16> %5, <16 x i16> zeroinitializer
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i16> %4, <16 x i16> zeroinitializer
%res3 = add <16 x i16> %3, %4
%res4 = add <16 x i16> %7, %res3
ret <16 x i16> %res4
ret <16 x i16> %res3
}
define <8 x i16> @test_int_x86_avx512_mask_vpshrdv_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16>* %x2p, <8 x i16> %x4, i8 %x3) {
@ -1261,10 +1221,7 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshrdv_w_128(<8 x i16> %x0, <8 x i16
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshrdvw (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x72,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshrdvw %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x72,0xe2]
; X86-NEXT: vpshrdvw %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x72,0xc2]
; X86-NEXT: vpaddw %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfd,0xc0]
; X86-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1273,10 +1230,7 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshrdv_w_128(<8 x i16> %x0, <8 x i16
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshrdvw (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x72,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshrdvw %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x72,0xe2]
; X64-NEXT: vpshrdvw %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x72,0xc2]
; X64-NEXT: vpaddw %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfd,0xc0]
; X64-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i16>, <8 x i16>* %x2p
@ -1284,12 +1238,10 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshrdv_w_128(<8 x i16> %x0, <8 x i16
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x0
%4 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> %x4)
%5 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i16> %5, <8 x i16> zeroinitializer
%res3 = add <8 x i16> %3, %4
%res4 = add <8 x i16> %7, %res3
ret <8 x i16> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i16> %4, <8 x i16> zeroinitializer
%res3 = add <8 x i16> %3, %6
ret <8 x i16> %res3
}
define <8 x i32> @test_int_x86_avx512_mask_vpshldv_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
@ -1300,10 +1252,7 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshldv_d_256(<8 x i32> %x0, <8 x i32
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshldvd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x71,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshldvd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x71,0xe2]
; X86-NEXT: vpshldvd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x71,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1312,10 +1261,7 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshldv_d_256(<8 x i32> %x0, <8 x i32
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshldvd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x71,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshldvd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x71,0xe2]
; X64-NEXT: vpshldvd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x71,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -1323,12 +1269,10 @@ define <8 x i32> @test_int_x86_avx512_mask_vpshldv_d_256(<8 x i32> %x0, <8 x i32
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
define <4 x i32> @test_int_x86_avx512_mask_vpshldv_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
@ -1339,10 +1283,7 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshldv_d_128(<4 x i32> %x0, <4 x i32
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshldvd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x71,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshldvd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x71,0xe2]
; X86-NEXT: vpshldvd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x71,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1351,10 +1292,7 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshldv_d_128(<4 x i32> %x0, <4 x i32
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshldvd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x71,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshldvd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x71,0xe2]
; X64-NEXT: vpshldvd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x71,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -1363,13 +1301,11 @@ define <4 x i32> @test_int_x86_avx512_mask_vpshldv_d_128(<4 x i32> %x0, <4 x i32
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}
define <4 x i64> @test_int_x86_avx512_mask_vpshldv_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64>* %x2p, <4 x i64> %x4, i8 %x3) {
@ -1380,10 +1316,7 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshldv_q_256(<4 x i64> %x0, <4 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshldvq (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x71,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshldvq %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x71,0xe2]
; X86-NEXT: vpshldvq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x71,0xc2]
; X86-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X86-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1392,10 +1325,7 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshldv_q_256(<4 x i64> %x0, <4 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshldvq (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x71,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshldvq %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x71,0xe2]
; X64-NEXT: vpshldvq %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x71,0xc2]
; X64-NEXT: vpaddq %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xd4,0xc0]
; X64-NEXT: vpaddq %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i64>, <4 x i64>* %x2p
@ -1404,13 +1334,11 @@ define <4 x i64> @test_int_x86_avx512_mask_vpshldv_q_256(<4 x i64> %x0, <4 x i64
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x0
%4 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x4)
%5 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i64> %5, <4 x i64> zeroinitializer
%res3 = add <4 x i64> %3, %4
%res4 = add <4 x i64> %7, %res3
ret <4 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i64> %4, <4 x i64> zeroinitializer
%res3 = add <4 x i64> %3, %6
ret <4 x i64> %res3
}
define <2 x i64> @test_int_x86_avx512_mask_vpshldv_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64>* %x2p, <2 x i64> %x4, i8 %x3) {
@ -1421,10 +1349,7 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshldv_q_128(<2 x i64> %x0, <2 x i64
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshldvq (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x71,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshldvq %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x71,0xe2]
; X86-NEXT: vpshldvq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x71,0xc2]
; X86-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X86-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1433,10 +1358,7 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshldv_q_128(<2 x i64> %x0, <2 x i64
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshldvq (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x71,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshldvq %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x71,0xe2]
; X64-NEXT: vpshldvq %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x71,0xc2]
; X64-NEXT: vpaddq %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xd4,0xc0]
; X64-NEXT: vpaddq %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xd4,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <2 x i64>, <2 x i64>* %x2p
@ -1445,13 +1367,11 @@ define <2 x i64> @test_int_x86_avx512_mask_vpshldv_q_128(<2 x i64> %x0, <2 x i64
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x0
%4 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x4)
%5 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <2 x i32> <i32 0, i32 1>
%7 = select <2 x i1> %extract1, <2 x i64> %5, <2 x i64> zeroinitializer
%res3 = add <2 x i64> %3, %4
%res4 = add <2 x i64> %7, %res3
ret <2 x i64> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract1, <2 x i64> %4, <2 x i64> zeroinitializer
%res3 = add <2 x i64> %3, %6
ret <2 x i64> %res3
}
define <16 x i16> @test_int_x86_avx512_mask_vpshldv_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16>* %x2p, <16 x i16> %x4, i16 %x3) {
@ -1461,10 +1381,7 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshldv_w_256(<16 x i16> %x0, <16 x
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpshldvw (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x70,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpshldvw %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x70,0xe2]
; X86-NEXT: vpshldvw %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x70,0xc2]
; X86-NEXT: vpaddw %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfd,0xc0]
; X86-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1473,10 +1390,7 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshldv_w_256(<16 x i16> %x0, <16 x
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpshldvw (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x29,0x70,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpshldvw %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0xf5,0x28,0x70,0xe2]
; X64-NEXT: vpshldvw %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0xa9,0x70,0xc2]
; X64-NEXT: vpaddw %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfd,0xc0]
; X64-NEXT: vpaddw %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i16>, <16 x i16>* %x2p
@ -1484,12 +1398,10 @@ define <16 x i16> @test_int_x86_avx512_mask_vpshldv_w_256(<16 x i16> %x0, <16 x
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i16> %1, <16 x i16> %x0
%4 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x4)
%5 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i16> %5, <16 x i16> zeroinitializer
%res3 = add <16 x i16> %3, %4
%res4 = add <16 x i16> %7, %res3
ret <16 x i16> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i16> %4, <16 x i16> zeroinitializer
%res3 = add <16 x i16> %3, %6
ret <16 x i16> %res3
}
define <8 x i16> @test_int_x86_avx512_mask_vpshldv_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16>* %x2p, <8 x i16> %x4, i8 %x3) {
@ -1500,10 +1412,7 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshldv_w_128(<8 x i16> %x0, <8 x i16
; X86-NEXT: kmovd %ecx, %k1 # encoding: [0xc5,0xfb,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpshldvw (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x70,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpshldvw %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x70,0xe2]
; X86-NEXT: vpshldvw %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x70,0xc2]
; X86-NEXT: vpaddw %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfd,0xc0]
; X86-NEXT: vpshldvw %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0x70,0xc2]
; X86-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -1512,10 +1421,7 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshldv_w_128(<8 x i16> %x0, <8 x i16
; X64-NEXT: kmovd %esi, %k1 # encoding: [0xc5,0xfb,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpshldvw (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0xf5,0x09,0x70,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpshldvw %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0xf5,0x08,0x70,0xe2]
; X64-NEXT: vpshldvw %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0xf5,0x89,0x70,0xc2]
; X64-NEXT: vpaddw %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfd,0xc0]
; X64-NEXT: vpshldvw %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0xf5,0x08,0x70,0xc2]
; X64-NEXT: vpaddw %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfd,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i16>, <8 x i16>* %x2p
@ -1523,12 +1429,10 @@ define <8 x i16> @test_int_x86_avx512_mask_vpshldv_w_128(<8 x i16> %x0, <8 x i16
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i16> %1, <8 x i16> %x0
%4 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x4)
%5 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i16> %5, <8 x i16> zeroinitializer
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i16> %4, <8 x i16> zeroinitializer
%res3 = add <8 x i16> %3, %4
%res4 = add <8 x i16> %7, %res3
ret <8 x i16> %res4
ret <8 x i16> %res3
}
declare <4 x i32> @llvm.fshl.v4i32(<4 x i32>, <4 x i32>, <4 x i32>)

2535
llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
View File

File diff suppressed because it is too large Load Diff

334
llvm/test/CodeGen/X86/avx512vl-intrinsics.ll
View File

@ -1538,102 +1538,122 @@ define <4 x i64>@test_int_x86_avx512_maskz_vpermt2var_q_256(<4 x i64> %x0, <4 x
declare <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)
define <2 x double>@test_int_x86_avx512_scalef_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2) {
; CHECK-LABEL: test_int_x86_avx512_scalef_pd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vscalefpd %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x2c,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
ret <2 x double> %res
}
define <2 x double>@test_int_x86_avx512_mask_scalef_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_scalef_pd_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vscalefpd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x2c,0xd1]
; X86-NEXT: vscalefpd %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x2c,0xc1]
; X86-NEXT: vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X86-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_scalef_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vscalefpd %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x09,0x2c,0xd1]
; X64-NEXT: vscalefpd %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0xfd,0x08,0x2c,0xc1]
; X64-NEXT: vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X64-NEXT: vmovapd %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
ret <2 x double> %res2
ret <2 x double> %res
}
declare <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)
define <4 x double>@test_int_x86_avx512_scalef_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2) {
; CHECK-LABEL: test_int_x86_avx512_scalef_pd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vscalefpd %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x2c,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
ret <4 x double> %res
}
define <4 x double>@test_int_x86_avx512_mask_scalef_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_scalef_pd_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vscalefpd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x2c,0xd1]
; X86-NEXT: vscalefpd %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x2c,0xc1]
; X86-NEXT: vaddpd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc0]
; X86-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_scalef_pd_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vscalefpd %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0xfd,0x29,0x2c,0xd1]
; X64-NEXT: vscalefpd %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0xfd,0x28,0x2c,0xc1]
; X64-NEXT: vaddpd %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0x58,0xc0]
; X64-NEXT: vmovapd %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
ret <4 x double> %res2
ret <4 x double> %res
}
declare <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
define <4 x float>@test_int_x86_avx512_scalef_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2) {
; CHECK-LABEL: test_int_x86_avx512_scalef_ps_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vscalefps %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0x2c,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
ret <4 x float> %res
}
define <4 x float>@test_int_x86_avx512_mask_scalef_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_scalef_ps_128:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vscalefps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x2c,0xd1]
; X86-NEXT: vscalefps %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0x2c,0xc1]
; X86-NEXT: vaddps %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe8,0x58,0xc0]
; X86-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_scalef_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vscalefps %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0x2c,0xd1]
; X64-NEXT: vscalefps %xmm1, %xmm0, %xmm0 # encoding: [0x62,0xf2,0x7d,0x08,0x2c,0xc1]
; X64-NEXT: vaddps %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe8,0x58,0xc0]
; X64-NEXT: vmovaps %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
ret <4 x float> %res2
ret <4 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
define <8 x float>@test_int_x86_avx512_scalef_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2) {
; CHECK-LABEL: test_int_x86_avx512_scalef_ps_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vscalefps %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0x7d,0x28,0x2c,0xc1]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
ret <8 x float> %res
}
define <8 x float>@test_int_x86_avx512_mask_scalef_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_scalef_ps_256:
; X86: # %bb.0:
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vscalefps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x2c,0xd1]
; X86-NEXT: vscalefps %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0x7d,0x28,0x2c,0xc1]
; X86-NEXT: vaddps %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc0]
; X86-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_scalef_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vscalefps %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0x2c,0xd1]
; X64-NEXT: vscalefps %ymm1, %ymm0, %ymm0 # encoding: [0x62,0xf2,0x7d,0x28,0x2c,0xc1]
; X64-NEXT: vaddps %ymm0, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xec,0x58,0xc0]
; X64-NEXT: vmovaps %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xc2]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
ret <8 x float> %res2
ret <8 x float> %res
}
declare <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64>, <16 x i8>, i8)
@ -4026,8 +4046,8 @@ define <2 x double>@test_int_x86_avx512_mask_getmant_pd_128(<2 x double> %x0, <2
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vgetmantpd $11, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x26,0xc8,0x0b]
; X86-NEXT: vgetmantpd $11, %xmm0, %xmm2 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0x26,0xd0,0x0b]
; X86-NEXT: vgetmantpd $11, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x26,0xc0,0x0b]
; X86-NEXT: vgetmantpd $12, %xmm0, %xmm2 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0x26,0xd0,0x0c]
; X86-NEXT: vgetmantpd $13, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x26,0xc0,0x0d]
; X86-NEXT: vaddpd %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x58,0xc0]
; X86-NEXT: vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
@ -4035,15 +4055,15 @@ define <2 x double>@test_int_x86_avx512_mask_getmant_pd_128(<2 x double> %x0, <2
; X64-LABEL: test_int_x86_avx512_mask_getmant_pd_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vgetmantpd $11, %xmm0, %xmm2 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0x26,0xd0,0x0b]
; X64-NEXT: vgetmantpd $11, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0x26,0xc8,0x0b]
; X64-NEXT: vgetmantpd $11, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x26,0xc0,0x0b]
; X64-NEXT: vgetmantpd $12, %xmm0, %xmm2 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0x26,0xd0,0x0c]
; X64-NEXT: vgetmantpd $13, %xmm0, %xmm0 # encoding: [0x62,0xf3,0xfd,0x08,0x26,0xc0,0x0d]
; X64-NEXT: vaddpd %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x58,0xc0]
; X64-NEXT: vaddpd %xmm0, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 %x3)
%res2 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> zeroinitializer, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 -1)
%res2 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 12, <2 x double> zeroinitializer, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 13, <2 x double> %x2, i8 -1)
%res3 = fadd <2 x double> %res, %res1
%res4 = fadd <2 x double> %res2, %res3
ret <2 x double> %res4
@ -4057,7 +4077,7 @@ define <4 x double>@test_int_x86_avx512_mask_getmant_pd_256(<4 x double> %x0, <4
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vgetmantpd $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x26,0xc8,0x0b]
; X86-NEXT: vgetmantpd $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x26,0xc0,0x0b]
; X86-NEXT: vgetmantpd $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x26,0xc0,0x0c]
; X86-NEXT: vaddpd %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0x58,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -4065,11 +4085,11 @@ define <4 x double>@test_int_x86_avx512_mask_getmant_pd_256(<4 x double> %x0, <4
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vgetmantpd $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0x26,0xc8,0x0b]
; X64-NEXT: vgetmantpd $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x26,0xc0,0x0b]
; X64-NEXT: vgetmantpd $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0xfd,0x28,0x26,0xc0,0x0c]
; X64-NEXT: vaddpd %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf5,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 -1)
%res1 = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 12, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
ret <4 x double> %res2
}
@ -4082,7 +4102,7 @@ define <4 x float>@test_int_x86_avx512_mask_getmant_ps_128(<4 x float> %x0, <4 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vgetmantps $11, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x26,0xc8,0x0b]
; X86-NEXT: vgetmantps $11, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x26,0xc0,0x0b]
; X86-NEXT: vgetmantps $12, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x26,0xc0,0x0c]
; X86-NEXT: vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -4090,11 +4110,11 @@ define <4 x float>@test_int_x86_avx512_mask_getmant_ps_128(<4 x float> %x0, <4 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vgetmantps $11, %xmm0, %xmm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x09,0x26,0xc8,0x0b]
; X64-NEXT: vgetmantps $11, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x26,0xc0,0x0b]
; X64-NEXT: vgetmantps $12, %xmm0, %xmm0 # encoding: [0x62,0xf3,0x7d,0x08,0x26,0xc0,0x0c]
; X64-NEXT: vaddps %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf0,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 -1)
%res1 = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 12, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
ret <4 x float> %res2
}
@ -4107,7 +4127,7 @@ define <8 x float>@test_int_x86_avx512_mask_getmant_ps_256(<8 x float> %x0, <8 x
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vgetmantps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x26,0xc8,0x0b]
; X86-NEXT: vgetmantps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x26,0xc0,0x0b]
; X86-NEXT: vgetmantps $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x26,0xc0,0x0c]
; X86-NEXT: vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -4115,43 +4135,44 @@ define <8 x float>@test_int_x86_avx512_mask_getmant_ps_256(<8 x float> %x0, <8 x
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vgetmantps $11, %ymm0, %ymm1 {%k1} # encoding: [0x62,0xf3,0x7d,0x29,0x26,0xc8,0x0b]
; X64-NEXT: vgetmantps $11, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x26,0xc0,0x0b]
; X64-NEXT: vgetmantps $12, %ymm0, %ymm0 # encoding: [0x62,0xf3,0x7d,0x28,0x26,0xc0,0x0c]
; X64-NEXT: vaddps %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xf4,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
%res1 = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 12, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
ret <8 x float> %res2
}
declare <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32)
define <4 x i32>@test_int_x86_avx512_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_d_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0x75,0x08,0x25,0xc2,0x21]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33)
ret <4 x i32> %1
}
define <4 x i32>@test_int_x86_avx512_mask_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_pternlog_d_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x25,0xc2,0x21]
; X86-NEXT: vpaddd %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfe,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pternlog_d_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x25,0xc2,0x21]
; X64-NEXT: vpaddd %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfe,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33)
%res2 = add <4 x i32> %3, %4
ret <4 x i32> %res2
ret <4 x i32> %3
}
declare <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32, i8)
@ -4159,58 +4180,51 @@ declare <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32>, <4 x i32>, <4
define <4 x i32>@test_int_x86_avx512_maskz_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_d_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x25,0xc2,0x21]
; X86-NEXT: vpaddd %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfe,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_d_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogd $33, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x25,0xc2,0x21]
; X64-NEXT: vpaddd %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xfe,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> zeroinitializer
%4 = call <4 x i32> @llvm.x86.avx512.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33)
%res2 = add <4 x i32> %3, %4
ret <4 x i32> %res2
ret <4 x i32> %3
}
declare <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32)
define <8 x i32>@test_int_x86_avx512_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_d_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0x75,0x28,0x25,0xc2,0x21]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33)
ret <8 x i32> %1
}
define <8 x i32>@test_int_x86_avx512_mask_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_pternlog_d_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x25,0xc2,0x21]
; X86-NEXT: vpaddd %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pternlog_d_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x25,0xc2,0x21]
; X64-NEXT: vpaddd %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
ret <8 x i32> %3
}
declare <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32, i8)
@ -4218,144 +4232,122 @@ declare <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32>, <8 x i32>, <8
define <8 x i32>@test_int_x86_avx512_maskz_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_d_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x25,0xc2,0x21]
; X86-NEXT: vpaddd %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_d_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogd $33, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x25,0xc2,0x21]
; X64-NEXT: vpaddd %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xfe,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> zeroinitializer
%4 = call <8 x i32> @llvm.x86.avx512.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33)
%res2 = add <8 x i32> %3, %4
ret <8 x i32> %res2
ret <8 x i32> %3
}
declare <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i32)
define <2 x i64>@test_int_x86_avx512_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_q_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0xf5,0x08,0x25,0xc2,0x21]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33)
ret <2 x i64> %1
}
define <2 x i64>@test_int_x86_avx512_mask_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_pternlog_q_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0xf5,0x08,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0xf5,0x09,0x25,0xc2,0x21]
; X86-NEXT: vpaddq %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xd4,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pternlog_q_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0xf5,0x08,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0xf5,0x09,0x25,0xc2,0x21]
; X64-NEXT: vpaddq %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xd4,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> %x0
%4 = call <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33)
%res2 = add <2 x i64> %3, %4
ret <2 x i64> %res2
ret <2 x i64> %3
}
define <2 x i64>@test_int_x86_avx512_maskz_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_q_128:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0xf5,0x08,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0xf5,0x89,0x25,0xc2,0x21]
; X86-NEXT: vpaddq %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xd4,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_q_128:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0xf5,0x08,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogq $33, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0xf5,0x89,0x25,0xc2,0x21]
; X64-NEXT: vpaddq %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xd4,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <2 x i32> <i32 0, i32 1>
%3 = select <2 x i1> %extract, <2 x i64> %1, <2 x i64> zeroinitializer
%4 = call <2 x i64> @llvm.x86.avx512.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33)
%res2 = add <2 x i64> %3, %4
ret <2 x i64> %res2
ret <2 x i64> %3
}
declare <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i32)
define <4 x i64>@test_int_x86_avx512_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: test_int_x86_avx512_pternlog_q_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0xf5,0x28,0x25,0xc2,0x21]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33)
ret <4 x i64> %1
}
define <4 x i64>@test_int_x86_avx512_mask_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_mask_pternlog_q_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0xf5,0x28,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0xf5,0x29,0x25,0xc2,0x21]
; X86-NEXT: vpaddq %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_pternlog_q_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0xf5,0x28,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0xf5,0x29,0x25,0xc2,0x21]
; X64-NEXT: vpaddq %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> %x0
%4 = call <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33)
%res2 = add <4 x i64> %3, %4
ret <4 x i64> %res2
ret <4 x i64> %3
}
define <4 x i64>@test_int_x86_avx512_maskz_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
; X86-LABEL: test_int_x86_avx512_maskz_pternlog_q_256:
; X86: # %bb.0:
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0xf5,0x28,0x25,0xda,0x21]
; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0xf5,0xa9,0x25,0xc2,0x21]
; X86-NEXT: vpaddq %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc3]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_pternlog_q_256:
; X64: # %bb.0:
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0xf5,0x28,0x25,0xda,0x21]
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vpternlogq $33, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0xf5,0xa9,0x25,0xc2,0x21]
; X64-NEXT: vpaddq %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xd4,0xc3]
; X64-NEXT: retq # encoding: [0xc3]
%1 = call <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33)
%2 = bitcast i8 %x4 to <8 x i1>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i64> %1, <4 x i64> zeroinitializer
%4 = call <4 x i64> @llvm.x86.avx512.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33)
%res2 = add <4 x i64> %3, %4
ret <4 x i64> %res2
ret <4 x i64> %3
}
define <8 x i16> @test_x86_vcvtps2ph_128(<4 x float> %a0, i8 %mask, <8 x i16> %src) {
@ -5047,29 +5039,29 @@ define <4 x float>@test_int_x86_avx512_mask_fixupimm_ps_128(<4 x float> %x0, <4
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xd8]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xda,0x05]
; X86-NEXT: vmovaps %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe0]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
; X86-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xc2,0x05]
; X86-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X86-NEXT: vmovaps %xmm0, %xmm5 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe8]
; X86-NEXT: vfixupimmps $6, %xmm4, %xmm1, %xmm5 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xec,0x06]
; X86-NEXT: vaddps %xmm5, %xmm3, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xdd]
; X86-NEXT: vfixupimmps $7, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xc2,0x07]
; X86-NEXT: vaddps %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xc0]
; X86-NEXT: vaddps %xmm4, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x58,0xc4]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_fixupimm_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xd8]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xda,0x05]
; X64-NEXT: vmovaps %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe0]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm4 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xe2,0x05]
; X64-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xc2,0x05]
; X64-NEXT: vaddps %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd8,0x58,0xc0]
; X64-NEXT: vaddps %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x58,0xc3]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xda,0x05]
; X64-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X64-NEXT: vmovaps %xmm0, %xmm5 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe8]
; X64-NEXT: vfixupimmps $6, %xmm4, %xmm1, %xmm5 {%k1} # encoding: [0x62,0xf3,0x75,0x09,0x54,0xec,0x06]
; X64-NEXT: vaddps %xmm5, %xmm3, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xdd]
; X64-NEXT: vfixupimmps $7, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xc2,0x07]
; X64-NEXT: vaddps %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 %x4)
%res1 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 5, i8 %x4)
%res2 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 -1)
%res1 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 6, i8 %x4)
%res2 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 7, i8 -1)
%res3 = fadd <4 x float> %res, %res1
%res4 = fadd <4 x float> %res3, %res2
ret <4 x float> %res4
@ -5084,29 +5076,29 @@ define <4 x float>@test_int_x86_avx512_maskz_fixupimm_ps_128(<4 x float> %x0, <4
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xd8]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xda,0x05]
; X86-NEXT: vmovaps %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe0]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
; X86-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X86-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xc2,0x05]
; X86-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X86-NEXT: vmovaps %xmm0, %xmm5 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe8]
; X86-NEXT: vfixupimmps $6, %xmm4, %xmm1, %xmm5 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xec,0x06]
; X86-NEXT: vaddps %xmm5, %xmm3, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xdd]
; X86-NEXT: vfixupimmps $7, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xc2,0x07]
; X86-NEXT: vaddps %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xc0]
; X86-NEXT: vaddps %xmm4, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x58,0xc4]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_128:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xd8]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xda,0x05]
; X64-NEXT: vmovaps %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe0]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xe2,0x05]
; X64-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xc2,0x05]
; X64-NEXT: vaddps %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd8,0x58,0xc0]
; X64-NEXT: vaddps %xmm3, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x58,0xc3]
; X64-NEXT: vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xda,0x05]
; X64-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X64-NEXT: vmovaps %xmm0, %xmm5 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0xe8]
; X64-NEXT: vfixupimmps $6, %xmm4, %xmm1, %xmm5 {%k1} {z} # encoding: [0x62,0xf3,0x75,0x89,0x54,0xec,0x06]
; X64-NEXT: vaddps %xmm5, %xmm3, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xdd]
; X64-NEXT: vfixupimmps $7, %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf3,0x75,0x08,0x54,0xc2,0x07]
; X64-NEXT: vaddps %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe0,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 %x4)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 5, i8 %x4)
%res2 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 -1)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 6, i8 %x4)
%res2 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 7, i8 -1)
%res3 = fadd <4 x float> %res, %res1
%res4 = fadd <4 x float> %res3, %res2
ret <4 x float> %res4
@ -5121,29 +5113,29 @@ define <8 x float>@test_int_x86_avx512_mask_fixupimm_ps_256(<8 x float> %x0, <8
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xd8]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xda,0x05]
; X86-NEXT: vmovaps %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe0]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
; X86-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xc2,0x05]
; X86-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X86-NEXT: vmovaps %ymm0, %ymm5 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe8]
; X86-NEXT: vfixupimmps $6, %ymm4, %ymm1, %ymm5 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xec,0x06]
; X86-NEXT: vaddps %ymm5, %ymm3, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xdd]
; X86-NEXT: vfixupimmps $7, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xc2,0x07]
; X86-NEXT: vaddps %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xc0]
; X86-NEXT: vaddps %ymm4, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc4]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_mask_fixupimm_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xd8]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xda,0x05]
; X64-NEXT: vmovaps %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe0]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm4 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xe2,0x05]
; X64-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xc2,0x05]
; X64-NEXT: vaddps %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdc,0x58,0xc0]
; X64-NEXT: vaddps %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc3]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xda,0x05]
; X64-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X64-NEXT: vmovaps %ymm0, %ymm5 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe8]
; X64-NEXT: vfixupimmps $6, %ymm4, %ymm1, %ymm5 {%k1} # encoding: [0x62,0xf3,0x75,0x29,0x54,0xec,0x06]
; X64-NEXT: vaddps %ymm5, %ymm3, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xdd]
; X64-NEXT: vfixupimmps $7, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xc2,0x07]
; X64-NEXT: vaddps %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 %x4)
%res1 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 5, i8 %x4)
%res2 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 -1)
%res1 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 6, i8 %x4)
%res2 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 7, i8 -1)
%res3 = fadd <8 x float> %res, %res1
%res4 = fadd <8 x float> %res3, %res2
ret <8 x float> %res4
@ -5158,29 +5150,29 @@ define <8 x float>@test_int_x86_avx512_maskz_fixupimm_ps_256(<8 x float> %x0, <8
; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
; X86-NEXT: vmovaps %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xd8]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xda,0x05]
; X86-NEXT: vmovaps %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe0]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
; X86-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X86-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xc2,0x05]
; X86-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X86-NEXT: vmovaps %ymm0, %ymm5 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe8]
; X86-NEXT: vfixupimmps $6, %ymm4, %ymm1, %ymm5 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xec,0x06]
; X86-NEXT: vaddps %ymm5, %ymm3, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xdd]
; X86-NEXT: vfixupimmps $7, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xc2,0x07]
; X86-NEXT: vaddps %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xc0]
; X86-NEXT: vaddps %ymm4, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc4]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_256:
; X64: # %bb.0:
; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
; X64-NEXT: vmovaps %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xd8]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xda,0x05]
; X64-NEXT: vmovaps %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe0]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xe2,0x05]
; X64-NEXT: vxorps %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe8,0x57,0xd2]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xc2,0x05]
; X64-NEXT: vaddps %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdc,0x58,0xc0]
; X64-NEXT: vaddps %ymm3, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x58,0xc3]
; X64-NEXT: vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xda,0x05]
; X64-NEXT: vxorps %xmm4, %xmm4, %xmm4 # encoding: [0xc5,0xd8,0x57,0xe4]
; X64-NEXT: vmovaps %ymm0, %ymm5 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x28,0xe8]
; X64-NEXT: vfixupimmps $6, %ymm4, %ymm1, %ymm5 {%k1} {z} # encoding: [0x62,0xf3,0x75,0xa9,0x54,0xec,0x06]
; X64-NEXT: vaddps %ymm5, %ymm3, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xdd]
; X64-NEXT: vfixupimmps $7, %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf3,0x75,0x28,0x54,0xc2,0x07]
; X64-NEXT: vaddps %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe4,0x58,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 %x4)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 5, i8 %x4)
%res2 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 -1)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 6, i8 %x4)
%res2 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 7, i8 -1)
%res3 = fadd <8 x float> %res, %res1
%res4 = fadd <8 x float> %res3, %res2
ret <8 x float> %res4

185
llvm/test/CodeGen/X86/avx512vl_vnni-intrinsics-upgrade.ll
View File

@ -5,6 +5,15 @@
declare <8 x i32> @llvm.x86.avx512.mask.vpdpbusd.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
declare <8 x i32> @llvm.x86.avx512.maskz.vpdpbusd.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
define <8 x i32>@test_int_x86_avx512_vpdpbusd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
ret <8 x i32> %res
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_256:
; X86: # %bb.0:
@ -13,10 +22,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x50,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpbusd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xe2]
; X86-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x50,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -25,24 +31,28 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x50,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpbusd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xe2]
; X64-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x50,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 -1)
%res2 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res3 = add <8 x i32> %res, %res1
%res4 = add <8 x i32> %res2, %res3
ret <8 x i32> %res4
%res1 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res2 = add <8 x i32> %res, %res1
ret <8 x i32> %res2
}
declare <4 x i32> @llvm.x86.avx512.mask.vpdpbusd.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
declare <4 x i32> @llvm.x86.avx512.maskz.vpdpbusd.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_vpdpbusd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_128:
; X86: # %bb.0:
@ -51,10 +61,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x50,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpbusd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xe2]
; X86-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x50,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -63,24 +70,28 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x50,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpbusd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xe2]
; X64-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x50,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 -1)
%res2 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res3 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
%res1 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res2 = add <4 x i32> %res, %res1
ret <4 x i32> %res2
}
declare <8 x i32> @llvm.x86.avx512.mask.vpdpbusds.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
declare <8 x i32> @llvm.x86.avx512.maskz.vpdpbusds.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
define <8 x i32>@test_int_x86_avx512_vpdpbusds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
ret <8 x i32> %res
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_256:
; X86: # %bb.0:
@ -89,10 +100,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x51,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpbusds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xe2]
; X86-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x51,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -101,24 +109,28 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x51,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpbusds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xe2]
; X64-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x51,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 -1)
%res2 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res3 = add <8 x i32> %res, %res1
%res4 = add <8 x i32> %res2, %res3
ret <8 x i32> %res4
%res1 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res2 = add <8 x i32> %res, %res1
ret <8 x i32> %res2
}
declare <4 x i32> @llvm.x86.avx512.mask.vpdpbusds.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
declare <4 x i32> @llvm.x86.avx512.maskz.vpdpbusds.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_vpdpbusds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_128:
; X86: # %bb.0:
@ -127,10 +139,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x51,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpbusds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xe2]
; X86-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x51,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -139,24 +148,28 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x51,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpbusds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xe2]
; X64-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x51,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 -1)
%res2 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res3 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
%res1 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res2 = add <4 x i32> %res, %res1
ret <4 x i32> %res2
}
declare <8 x i32> @llvm.x86.avx512.mask.vpdpwssd.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
declare <8 x i32> @llvm.x86.avx512.maskz.vpdpwssd.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
define <8 x i32>@test_int_x86_avx512_vpdpwssd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
ret <8 x i32> %res
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_256:
; X86: # %bb.0:
@ -165,10 +178,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x52,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpwssd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xe2]
; X86-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x52,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -177,24 +187,28 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x52,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpwssd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xe2]
; X64-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x52,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 -1)
%res2 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res3 = add <8 x i32> %res, %res1
%res4 = add <8 x i32> %res2, %res3
ret <8 x i32> %res4
%res1 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res2 = add <8 x i32> %res, %res1
ret <8 x i32> %res2
}
declare <4 x i32> @llvm.x86.avx512.mask.vpdpwssd.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
declare <4 x i32> @llvm.x86.avx512.maskz.vpdpwssd.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_vpdpwssd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_128:
; X86: # %bb.0:
@ -203,10 +217,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x52,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpwssd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xe2]
; X86-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x52,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -215,25 +226,29 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x52,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpwssd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xe2]
; X64-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x52,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 -1)
%res2 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res3 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
%res1 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res2 = add <4 x i32> %res, %res1
ret <4 x i32> %res2
}
declare <8 x i32> @llvm.x86.avx512.mask.vpdpwssds.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
declare <8 x i32> @llvm.x86.avx512.maskz.vpdpwssds.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)
define <8 x i32>@test_int_x86_avx512_vpdpwssds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssds_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
ret <8 x i32> %res
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_256:
; X86: # %bb.0:
@ -242,10 +257,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x53,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpwssds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xe2]
; X86-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x53,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -254,24 +266,28 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x53,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpwssds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xe2]
; X64-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x53,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
%res = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 -1)
%res2 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res3 = add <8 x i32> %res, %res1
%res4 = add <8 x i32> %res2, %res3
ret <8 x i32> %res4
%res1 = call <8 x i32> @llvm.x86.avx512.maskz.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4, i8 %x3)
%res2 = add <8 x i32> %res, %res1
ret <8 x i32> %res2
}
declare <4 x i32> @llvm.x86.avx512.mask.vpdpwssds.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
declare <4 x i32> @llvm.x86.avx512.maskz.vpdpwssds.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_vpdpwssds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssds_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssds %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x53,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
ret <4 x i32> %res
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_128:
; X86: # %bb.0:
@ -280,10 +296,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x53,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpwssds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x53,0xe2]
; X86-NEXT: vpdpwssds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x53,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -292,18 +305,12 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x53,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpwssds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x53,0xe2]
; X64-NEXT: vpdpwssds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x53,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
%res = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 -1)
%res2 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res3 = add <4 x i32> %res, %res1
%res4 = add <4 x i32> %res2, %res3
ret <4 x i32> %res4
%res1 = call <4 x i32> @llvm.x86.avx512.maskz.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4, i8 %x3)
%res2 = add <4 x i32> %res, %res1
ret <4 x i32> %res2
}

216
llvm/test/CodeGen/X86/avx512vl_vnni-intrinsics.ll
View File

@ -4,6 +4,15 @@
declare <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32>, <8 x i32>, <8 x i32>)
define <8 x i32>@test_int_x86_avx512_vpdpbusd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2)
ret <8 x i32> %1
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_256:
; X86: # %bb.0:
@ -12,10 +21,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x50,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpbusd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xe2]
; X86-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x50,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -24,10 +30,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x50,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpbusd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x50,0xe2]
; X64-NEXT: vpdpbusd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x50,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -35,16 +38,23 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusd_256(<8 x i32> %x0, <8 x i32>
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.x86.avx512.vpdpbusd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
declare <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32>, <4 x i32>, <4 x i32>)
define <4 x i32>@test_int_x86_avx512_vpdpbusd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2)
ret <4 x i32> %1
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_128:
; X86: # %bb.0:
@ -53,10 +63,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x50,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpbusd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xe2]
; X86-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x50,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -65,10 +72,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x50,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpbusd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x50,0xe2]
; X64-NEXT: vpdpbusd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x50,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -77,17 +81,24 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusd_128(<4 x i32> %x0, <4 x i32>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.x86.avx512.vpdpbusd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}
declare <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32>, <8 x i32>, <8 x i32>)
define <8 x i32>@test_int_x86_avx512_vpdpbusds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2)
ret <8 x i32> %1
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_256:
; X86: # %bb.0:
@ -96,10 +107,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x51,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpbusds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xe2]
; X86-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x51,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -108,10 +116,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x51,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpbusds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x51,0xe2]
; X64-NEXT: vpdpbusds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x51,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -119,16 +124,23 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpbusds_256(<8 x i32> %x0, <8 x i32>
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.x86.avx512.vpdpbusds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
declare <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32>, <4 x i32>, <4 x i32>)
define <4 x i32>@test_int_x86_avx512_vpdpbusds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2)
ret <4 x i32> %1
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_128:
; X86: # %bb.0:
@ -137,10 +149,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x51,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpbusds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xe2]
; X86-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x51,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -149,10 +158,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x51,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpbusds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x51,0xe2]
; X64-NEXT: vpdpbusds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x51,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -161,17 +167,24 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpbusds_128(<4 x i32> %x0, <4 x i32>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.x86.avx512.vpdpbusds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}
declare <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32>, <8 x i32>, <8 x i32>)
define <8 x i32>@test_int_x86_avx512_vpdpwssd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2)
ret <8 x i32> %1
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_256:
; X86: # %bb.0:
@ -180,10 +193,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x52,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpwssd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xe2]
; X86-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x52,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -192,10 +202,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x52,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpwssd %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x52,0xe2]
; X64-NEXT: vpdpwssd %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x52,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -203,16 +210,23 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssd_256(<8 x i32> %x0, <8 x i32>
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.x86.avx512.vpdpwssd.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
declare <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32>, <4 x i32>, <4 x i32>)
define <4 x i32>@test_int_x86_avx512_vpdpwssd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2)
ret <4 x i32> %1
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_128:
; X86: # %bb.0:
@ -221,10 +235,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x52,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpwssd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xe2]
; X86-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x52,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -233,10 +244,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x52,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpwssd %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x52,0xe2]
; X64-NEXT: vpdpwssd %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x52,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -245,17 +253,24 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssd_128(<4 x i32> %x0, <4 x i32>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.x86.avx512.vpdpwssd.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}
declare <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32>, <8 x i32>, <8 x i32>)
define <8 x i32>@test_int_x86_avx512_vpdpwssds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssds_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2)
ret <8 x i32> %1
}
define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>* %x2p, <8 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_256:
; X86: # %bb.0:
@ -264,10 +279,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x53,0x18]
; X86-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X86-NEXT: vpdpwssds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xe2]
; X86-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x53,0xc2]
; X86-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X86-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -276,10 +288,7 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %ymm1, %ymm3 {%k1} # encoding: [0x62,0xf2,0x75,0x29,0x53,0x1f]
; X64-NEXT: vmovdqa %ymm0, %ymm4 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x6f,0xe0]
; X64-NEXT: vpdpwssds %ymm2, %ymm1, %ymm4 # encoding: [0x62,0xf2,0x75,0x28,0x53,0xe2]
; X64-NEXT: vpdpwssds %ymm2, %ymm1, %ymm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xa9,0x53,0xc2]
; X64-NEXT: vpaddd %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xfe,0xc0]
; X64-NEXT: vpaddd %ymm0, %ymm3, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xe5,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <8 x i32>, <8 x i32>* %x2p
@ -287,16 +296,30 @@ define <8 x i32>@test_int_x86_avx512_mask_vpdpwssds_256(<8 x i32> %x0, <8 x i32>
%2 = bitcast i8 %x3 to <8 x i1>
%3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %x0
%4 = call <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%5 = call <8 x i32> @llvm.x86.avx512.vpdpwssds.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%7 = select <8 x i1> %6, <8 x i32> %5, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %4
%res4 = add <8 x i32> %7, %res3
ret <8 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%6 = select <8 x i1> %5, <8 x i32> %4, <8 x i32> zeroinitializer
%res3 = add <8 x i32> %3, %6
ret <8 x i32> %res3
}
declare <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32>, <4 x i32>, <4 x i32>)
define <4 x i32>@test_int_x86_avx512_vpdpwssds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p) {
; X86-LABEL: test_int_x86_avx512_vpdpwssds_128:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X86-NEXT: vpdpwssds (%eax), %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x53,0x00]
; X86-NEXT: retl # encoding: [0xc3]
;
; X64-LABEL: test_int_x86_avx512_vpdpwssds_128:
; X64: # %bb.0:
; X64-NEXT: vpdpwssds (%rdi), %xmm1, %xmm0 # encoding: [0x62,0xf2,0x75,0x08,0x53,0x07]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
%1 = call <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2)
ret <4 x i32> %1
}
define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>* %x2p, <4 x i32> %x4, i8 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_128:
; X86: # %bb.0:
@ -305,10 +328,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32>
; X86-NEXT: kmovw %ecx, %k1 # encoding: [0xc5,0xf8,0x92,0xc9]
; X86-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x53,0x18]
; X86-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X86-NEXT: vpdpwssds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x53,0xe2]
; X86-NEXT: vpdpwssds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x53,0xc2]
; X86-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X86-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -317,10 +337,7 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32>
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %xmm1, %xmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x09,0x53,0x1f]
; X64-NEXT: vmovdqa %xmm0, %xmm4 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x6f,0xe0]
; X64-NEXT: vpdpwssds %xmm2, %xmm1, %xmm4 # encoding: [0x62,0xf2,0x75,0x08,0x53,0xe2]
; X64-NEXT: vpdpwssds %xmm2, %xmm1, %xmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0x89,0x53,0xc2]
; X64-NEXT: vpaddd %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xfe,0xc0]
; X64-NEXT: vpaddd %xmm0, %xmm3, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe1,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <4 x i32>, <4 x i32>* %x2p
@ -329,12 +346,9 @@ define <4 x i32>@test_int_x86_avx512_mask_vpdpwssds_128(<4 x i32> %x0, <4 x i32>
%extract = shufflevector <8 x i1> %2, <8 x i1> %2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = select <4 x i1> %extract, <4 x i32> %1, <4 x i32> %x0
%4 = call <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%5 = call <4 x i32> @llvm.x86.avx512.vpdpwssds.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x4)
%6 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %6, <8 x i1> %6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%7 = select <4 x i1> %extract1, <4 x i32> %5, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %4
%res4 = add <4 x i32> %7, %res3
ret <4 x i32> %res4
%5 = bitcast i8 %x3 to <8 x i1>
%extract1 = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract1, <4 x i32> %4, <4 x i32> zeroinitializer
%res3 = add <4 x i32> %3, %6
ret <4 x i32> %res3
}

92
llvm/test/CodeGen/X86/avx512vnni-intrinsics-upgrade.ll
View File

@ -5,6 +5,15 @@
declare <16 x i32> @llvm.x86.avx512.mask.vpdpbusd.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpdpbusd.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpdpbusd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusd_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_512:
; X86: # %bb.0:
@ -12,10 +21,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x50,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpbusd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xe2]
; X86-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x50,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -24,24 +30,28 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i3
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x50,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpbusd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xe2]
; X64-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x50,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
declare <16 x i32> @llvm.x86.avx512.mask.vpdpbusds.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpdpbusds.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpdpbusds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_512:
; X86: # %bb.0:
@ -49,10 +59,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x51,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpbusds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xe2]
; X86-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x51,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -61,24 +68,28 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x51,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpbusds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xe2]
; X64-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x51,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
declare <16 x i32> @llvm.x86.avx512.mask.vpdpwssd.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpdpwssd.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpdpwssd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_512:
; X86: # %bb.0:
@ -86,10 +97,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x52,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpwssd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xe2]
; X86-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x52,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -98,24 +106,28 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i3
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x52,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpwssd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xe2]
; X64-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x52,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}
declare <16 x i32> @llvm.x86.avx512.mask.vpdpwssds.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
declare <16 x i32> @llvm.x86.avx512.maskz.vpdpwssds.512(<16 x i32>, <16 x i32>, <16 x i32>, i16)
define <16 x i32>@test_int_x86_avx512_vpdpwssds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssds_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 -1)
ret <16 x i32> %res
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_512:
; X86: # %bb.0:
@ -123,10 +135,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x53,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpwssds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xe2]
; X86-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x53,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -135,18 +144,13 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x53,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpwssds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xe2]
; X64-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x53,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
%res = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, i16 %x3)
%res1 = call <16 x i32> @llvm.x86.avx512.mask.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 -1)
%res2 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res3 = add <16 x i32> %res, %res1
%res4 = add <16 x i32> %res2, %res3
ret <16 x i32> %res4
%res1 = call <16 x i32> @llvm.x86.avx512.maskz.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4, i16 %x3)
%res2 = add <16 x i32> %res, %res1
ret <16 x i32> %res2
}

101
llvm/test/CodeGen/X86/avx512vnni-intrinsics.ll
View File

@ -4,6 +4,15 @@
declare <16 x i32> @llvm.x86.avx512.vpdpbusd.512(<16 x i32>, <16 x i32>, <16 x i32>)
define <16 x i32>@test_int_x86_avx512_ask_vpdpbusd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_ask_vpdpbusd_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2)
ret <16 x i32> %1
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusd_512:
; X86: # %bb.0:
@ -11,10 +20,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpbusd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x50,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpbusd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xe2]
; X86-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x50,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -23,10 +29,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i3
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpbusd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x50,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpbusd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x50,0xe2]
; X64-NEXT: vpdpbusd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x50,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -34,16 +37,23 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusd_512(<16 x i32> %x0, <16 x i3
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.x86.avx512.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.x86.avx512.vpdpbusd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}
declare <16 x i32> @llvm.x86.avx512.vpdpbusds.512(<16 x i32>, <16 x i32>, <16 x i32>)
define <16 x i32>@test_int_x86_avx512_vpdpbusds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpbusds_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2)
ret <16 x i32> %1
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpbusds_512:
; X86: # %bb.0:
@ -51,10 +61,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpbusds (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x51,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpbusds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xe2]
; X86-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x51,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -63,10 +70,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpbusds (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x51,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpbusds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x51,0xe2]
; X64-NEXT: vpdpbusds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x51,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -74,16 +78,23 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpbusds_512(<16 x i32> %x0, <16 x i
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.x86.avx512.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.x86.avx512.vpdpbusds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}
declare <16 x i32> @llvm.x86.avx512.vpdpwssd.512(<16 x i32>, <16 x i32>, <16 x i32>)
define <16 x i32>@test_int_x86_avx512_vpdpwssd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_vpdpwssd_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2)
ret <16 x i32> %1
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssd_512:
; X86: # %bb.0:
@ -91,10 +102,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i3
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpwssd (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x52,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpwssd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xe2]
; X86-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x52,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -103,10 +111,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i3
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpwssd (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x52,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpwssd %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x52,0xe2]
; X64-NEXT: vpdpwssd %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x52,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -114,16 +119,23 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssd_512(<16 x i32> %x0, <16 x i3
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.x86.avx512.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.x86.avx512.vpdpwssd.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}
declare <16 x i32> @llvm.x86.avx512.vpdpwssds.512(<16 x i32>, <16 x i32>, <16 x i32>)
define <16 x i32>@test_int_x86_avx512_ask_vpdpwssds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2) {
; CHECK-LABEL: test_int_x86_avx512_ask_vpdpwssds_512:
; CHECK: # %bb.0:
; CHECK-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xc2]
; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%1 = call <16 x i32> @llvm.x86.avx512.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2)
ret <16 x i32> %1
}
define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32>* %x2p, <16 x i32> %x4, i16 %x3) {
; X86-LABEL: test_int_x86_avx512_mask_vpdpwssds_512:
; X86: # %bb.0:
@ -131,10 +143,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i
; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x08]
; X86-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X86-NEXT: vpdpwssds (%eax), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x53,0x18]
; X86-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X86-NEXT: vpdpwssds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xe2]
; X86-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x53,0xc2]
; X86-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X86-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X86-NEXT: retl # encoding: [0xc3]
;
@ -143,10 +152,7 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i
; X64-NEXT: kmovw %esi, %k1 # encoding: [0xc5,0xf8,0x92,0xce]
; X64-NEXT: vmovdqa64 %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xd8]
; X64-NEXT: vpdpwssds (%rdi), %zmm1, %zmm3 {%k1} # encoding: [0x62,0xf2,0x75,0x49,0x53,0x1f]
; X64-NEXT: vmovdqa64 %zmm0, %zmm4 # encoding: [0x62,0xf1,0xfd,0x48,0x6f,0xe0]
; X64-NEXT: vpdpwssds %zmm2, %zmm1, %zmm4 # encoding: [0x62,0xf2,0x75,0x48,0x53,0xe2]
; X64-NEXT: vpdpwssds %zmm2, %zmm1, %zmm0 {%k1} {z} # encoding: [0x62,0xf2,0x75,0xc9,0x53,0xc2]
; X64-NEXT: vpaddd %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0x5d,0x48,0xfe,0xc0]
; X64-NEXT: vpaddd %zmm0, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x65,0x48,0xfe,0xc0]
; X64-NEXT: retq # encoding: [0xc3]
%x2 = load <16 x i32>, <16 x i32>* %x2p
@ -154,11 +160,8 @@ define <16 x i32>@test_int_x86_avx512_mask_vpdpwssds_512(<16 x i32> %x0, <16 x i
%2 = bitcast i16 %x3 to <16 x i1>
%3 = select <16 x i1> %2, <16 x i32> %1, <16 x i32> %x0
%4 = call <16 x i32> @llvm.x86.avx512.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%5 = call <16 x i32> @llvm.x86.avx512.vpdpwssds.512(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x4)
%6 = bitcast i16 %x3 to <16 x i1>
%7 = select <16 x i1> %6, <16 x i32> %5, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %4
%res4 = add <16 x i32> %7, %res3
ret <16 x i32> %res4
%5 = bitcast i16 %x3 to <16 x i1>
%6 = select <16 x i1> %5, <16 x i32> %4, <16 x i32> zeroinitializer
%res3 = add <16 x i32> %3, %6
ret <16 x i32> %res3
}