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

[X86][XOP] More VPPERM shuffle mask decode tests 

As requested by D18441

llvm-svn: 266531
This commit is contained in:
Simon Pilgrim 2016-04-16 16:37:21 +00:00
parent 37a09e72bc
commit 4eb94a93a0
1 changed files with 104 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

104
llvm/test/CodeGen/X86/xop-mask-comments.ll Normal file
View File

@ -0,0 +1,104 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx,+xop | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+xop | FileCheck %s --check-prefix=X64
;
; VPPERM
;
define <16 x i8> @vpperm_shuffle_unary(<16 x i8> %a0) {
; X32-LABEL: vpperm_shuffle_unary:
; X32: # BB#0:
; X32-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_unary:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a0, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 19, i8 2, i8 17, i8 0>)
ret <16 x i8> %1
}
define <16 x i8> @vpperm_shuffle_unary_undef(<16 x i8> %a0) {
; X32-LABEL: vpperm_shuffle_unary_undef:
; X32: # BB#0:
; X32-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_unary_undef:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,0]
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> undef, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 19, i8 2, i8 17, i8 0>)
ret <16 x i8> %1
}
define <16 x i8> @vpperm_shuffle_unary_zero(<16 x i8> %a0) {
; X32-LABEL: vpperm_shuffle_unary_zero:
; X32: # BB#0:
; X32-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3],zero,xmm0[1],zero
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_unary_zero:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*#+}} xmm0 = xmm0[15,14,13,12,11,10,9,8,7,6,5,4,3],zero,xmm0[1],zero
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a0, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 19, i8 130, i8 17, i8 128>)
ret <16 x i8> %1
}
define <16 x i8> @vpperm_shuffle_binary(<16 x i8> %a0, <16 x i8> %a1) {
; X32-LABEL: vpperm_shuffle_binary:
; X32: # BB#0:
; X32-NEXT: vpperm {{.*#+}} xmm0 = xmm1[15],xmm0[14],xmm1[13],xmm0[12],xmm1[11],xmm0[10],xmm1[9],xmm0[8],xmm1[7],xmm0[6],xmm1[5],xmm0[4],xmm1[3],xmm0[2],xmm1[1],xmm0[0]
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_binary:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*#+}} xmm0 = xmm1[15],xmm0[14],xmm1[13],xmm0[12],xmm1[11],xmm0[10],xmm1[9],xmm0[8],xmm1[7],xmm0[6],xmm1[5],xmm0[4],xmm1[3],xmm0[2],xmm1[1],xmm0[0]
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 19, i8 2, i8 17, i8 0>)
ret <16 x i8> %1
}
define <16 x i8> @vpperm_shuffle_binary_zero(<16 x i8> %a0, <16 x i8> %a1) {
; X32-LABEL: vpperm_shuffle_binary_zero:
; X32: # BB#0:
; X32-NEXT: vpperm {{.*#+}} xmm0 = xmm1[15],xmm0[14],xmm1[13],xmm0[12],xmm1[11],xmm0[10],xmm1[9],xmm0[8],xmm1[7],xmm0[6],xmm1[5],xmm0[4],zero,zero,zero,zero
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_binary_zero:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*#+}} xmm0 = xmm1[15],xmm0[14],xmm1[13],xmm0[12],xmm1[11],xmm0[10],xmm1[9],xmm0[8],xmm1[7],xmm0[6],xmm1[5],xmm0[4],zero,zero,zero,zero
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a1, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 147, i8 130, i8 145, i8 128>)
ret <16 x i8> %1
}
; we can't decode vpperm's other permute ops
define <16 x i8> @vpperm_shuffle_general(<16 x i8> %a0, <16 x i8> %a1) {
; X32-LABEL: vpperm_shuffle_general:
; X32: # BB#0:
; X32-NEXT: vpperm .LCPI5_0, %xmm0, %xmm0, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: vpperm_shuffle_general:
; X64: # BB#0:
; X64-NEXT: vpperm {{.*}}(%rip), %xmm0, %xmm0, %xmm0
; X64-NEXT: retq
%1 = tail call <16 x i8> @llvm.x86.xop.vpperm(<16 x i8> %a0, <16 x i8> %a0, <16 x i8> <i8 31, i8 14, i8 29, i8 12, i8 27, i8 10, i8 25, i8 8, i8 23, i8 6, i8 21, i8 4, i8 179, i8 162, i8 177, i8 160>)
ret <16 x i8> %1
}
;
; VPERMIL2
;
declare <2 x double> @llvm.x86.xop.vpermil2pd(<2 x double>, <2 x double>, <2 x double>, i8) nounwind readnone
declare <4 x double> @llvm.x86.xop.vpermil2pd.256(<4 x double>, <4 x double>, <4 x double>, i8) nounwind readnone
declare <4 x float> @llvm.x86.xop.vpermil2ps(<4 x float>, <4 x float>, <4 x float>, i8) nounwind readnone
declare <8 x float> @llvm.x86.xop.vpermil2ps.256(<8 x float>, <8 x float>, <8 x float>, i8) nounwind readnone
declare <16 x i8> @llvm.x86.xop.vpperm(<16 x i8>, <16 x i8>, <16 x i8>) nounwind readnone