maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/vector-trunc-widen.ll

2127 lines
80 KiB
LLVM
Raw Normal View History

[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE2
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+ssse3 | FileCheck %s --check-prefix=SSE --check-prefix=SSSE3
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE --check-prefix=SSE41
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2,AVX2-SLOW
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx2,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX2-FAST
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX512 --check-prefix=AVX512F
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512VL
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512bw,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512BW
; RUN: llc < %s -x86-experimental-vector-widening-legalization -mtriple=x86_64-unknown-unknown -mattr=+avx512bw,+avx512vl,+fast-variable-shuffle | FileCheck %s --check-prefixes=AVX512,AVX512BWVL
define <8 x i32> @trunc8i64_8i32(<8 x i64> %a) {
; SSE-LABEL: trunc8i64_8i32:
; SSE: # %bb.0: # %entry
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
; SSE-NEXT: movaps %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc8i64_8i32:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc8i64_8i32:
; AVX2-SLOW: # %bb.0: # %entry
[x86] split more v8f32/v8i32 shuffles in lowering Similar to D57867 - this is a small patch with lots of test diffs. With half-vector-width narrowing potential, using an extract + 128-bit vshufps is a win because it replaces a 256-bit shuffle with a 128-bit shufle. This seems like it should be a win even for targets with 'fast-variable-shuffle', but we are intentionally deferring that to an independent change to make sure that is true. Differential Revision: https://reviews.llvm.org/D58181 llvm-svn: 354279
2019-02-19 00:46:12 +08:00
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc8i64_8i32:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT: vpermps %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-FAST-NEXT: retq
;
; AVX512-LABEL: trunc8i64_8i32:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpmovqd %zmm0, %ymm0
; AVX512-NEXT: retq
entry:
%0 = trunc <8 x i64> %a to <8 x i32>
ret <8 x i32> %0
}
define <8 x i32> @trunc8i64_8i32_ashr(<8 x i64> %a) {
; SSE-LABEL: trunc8i64_8i32_ashr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
; SSE-NEXT: movaps %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc8i64_8i32_ashr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
; AVX1-NEXT: vshufps {{.*#+}} xmm1 = xmm1[1,3],xmm3[1,3]
; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[1,3]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc8i64_8i32_ashr:
; AVX2-SLOW: # %bb.0: # %entry
[x86] split more v8f32/v8i32 shuffles in lowering Similar to D57867 - this is a small patch with lots of test diffs. With half-vector-width narrowing potential, using an extract + 128-bit vshufps is a win because it replaces a 256-bit shuffle with a 128-bit shufle. This seems like it should be a win even for targets with 'fast-variable-shuffle', but we are intentionally deferring that to an independent change to make sure that is true. Differential Revision: https://reviews.llvm.org/D58181 llvm-svn: 354279
2019-02-19 00:46:12 +08:00
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[1,3],xmm2[1,3]
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[1,3]
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc8i64_8i32_ashr:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = [1,3,5,7,5,7,6,7]
; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT: vpermps %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-FAST-NEXT: retq
;
; AVX512-LABEL: trunc8i64_8i32_ashr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsraq $32, %zmm0, %zmm0
; AVX512-NEXT: vpmovqd %zmm0, %ymm0
; AVX512-NEXT: retq
entry:
%0 = ashr <8 x i64> %a, <i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32>
%1 = trunc <8 x i64> %0 to <8 x i32>
ret <8 x i32> %1
}
define <8 x i32> @trunc8i64_8i32_lshr(<8 x i64> %a) {
; SSE-LABEL: trunc8i64_8i32_lshr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[1,3],xmm3[1,3]
; SSE-NEXT: movaps %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc8i64_8i32_lshr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
; AVX1-NEXT: vshufps {{.*#+}} xmm1 = xmm1[1,3],xmm3[1,3]
; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[1,3]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc8i64_8i32_lshr:
; AVX2-SLOW: # %bb.0: # %entry
; AVX2-SLOW-NEXT: vpsrlq $32, %ymm0, %ymm0
[x86] split more v8f32/v8i32 shuffles in lowering Similar to D57867 - this is a small patch with lots of test diffs. With half-vector-width narrowing potential, using an extract + 128-bit vshufps is a win because it replaces a 256-bit shuffle with a 128-bit shufle. This seems like it should be a win even for targets with 'fast-variable-shuffle', but we are intentionally deferring that to an independent change to make sure that is true. Differential Revision: https://reviews.llvm.org/D58181 llvm-svn: 354279
2019-02-19 00:46:12 +08:00
; AVX2-SLOW-NEXT: vpsrlq $32, %ymm1, %ymm1
; AVX2-SLOW-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vextracti128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc8i64_8i32_lshr:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vpsrlq $32, %ymm1, %ymm1
; AVX2-FAST-NEXT: vpsrlq $32, %ymm0, %ymm0
; AVX2-FAST-NEXT: vmovdqa {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
; AVX2-FAST-NEXT: vpermd %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT: vpermd %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX2-FAST-NEXT: retq
;
; AVX512-LABEL: trunc8i64_8i32_lshr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsrlq $32, %zmm0, %zmm0
; AVX512-NEXT: vpmovqd %zmm0, %ymm0
; AVX512-NEXT: retq
entry:
%0 = lshr <8 x i64> %a, <i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32>
%1 = trunc <8 x i64> %0 to <8 x i32>
ret <8 x i32> %1
}
define <8 x i16> @trunc8i64_8i16(<8 x i64> %a) {
; SSE2-LABEL: trunc8i64_8i16:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
; SSE2-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i64_8i16:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
; SSSE3-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSSE3-NEXT: movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i64_8i16:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: pxor %xmm4, %xmm4
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm4[1,2,3],xmm3[4],xmm4[5,6,7]
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0],xmm4[1,2,3],xmm2[4],xmm4[5,6,7]
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1,2,3],xmm1[4],xmm4[5,6,7]
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1,2,3],xmm0[4],xmm4[5,6,7]
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: packusdw %xmm2, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc8i64_8i16:
; AVX1: # %bb.0: # %entry
[X86] combineVectorTruncationWithPACKUS - remove split/concatenation of mask combineVectorTruncationWithPACKUS is currently splitting the upper bit bit masking into 128-bit subregs and then concatenating them back together. This was originally done to avoid regressions that caused existing subregs to be concatenated to the larger type just for the AND masking before being extracted again. This was fixed by @spatel (notably rL303997 and rL347356). This also lets SimplifyDemandedBits do some further improvements before it hits the recursive depth limit. My only annoyance with this is that we were broadcasting some xmm masks but we seem to have lost them by moving to ymm - but that's a known issue as the logic in lowerBuildVectorAsBroadcast isn't great. Differential Revision: https://reviews.llvm.org/D60375#inline-539623 llvm-svn: 358692
2019-04-19 01:23:09 +08:00
; AVX1-NEXT: vmovaps {{.*#+}} ymm2 = [65535,65535,65535,65535]
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
; AVX1-NEXT: vpackusdw %xmm3, %xmm1, %xmm1
; AVX1-NEXT: vandps %ymm2, %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc8i64_8i16:
; AVX2-SLOW: # %bb.0: # %entry
[x86] split more v8f32/v8i32 shuffles in lowering Similar to D57867 - this is a small patch with lots of test diffs. With half-vector-width narrowing potential, using an extract + 128-bit vshufps is a win because it replaces a 256-bit shuffle with a 128-bit shufle. This seems like it should be a win even for targets with 'fast-variable-shuffle', but we are intentionally deferring that to an independent change to make sure that is true. Differential Revision: https://reviews.llvm.org/D58181 llvm-svn: 354279
2019-02-19 00:46:12 +08:00
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-SLOW-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
; AVX2-SLOW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX2-SLOW-NEXT: vzeroupper
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc8i64_8i16:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vmovdqa {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
; AVX2-FAST-NEXT: vpermd %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT: vpermd %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX2-FAST-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-FAST-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
; AVX2-FAST-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX2-FAST-NEXT: vzeroupper
; AVX2-FAST-NEXT: retq
;
; AVX512-LABEL: trunc8i64_8i16:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpmovqw %zmm0, %xmm0
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = trunc <8 x i64> %a to <8 x i16>
ret <8 x i16> %0
}
define void @trunc8i64_8i8(<8 x i64> %a) {
; SSE2-LABEL: trunc8i64_8i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0]
; SSE2-NEXT: pand %xmm4, %xmm3
; SSE2-NEXT: pand %xmm4, %xmm2
; SSE2-NEXT: packuswb %xmm3, %xmm2
; SSE2-NEXT: pand %xmm4, %xmm1
; SSE2-NEXT: pand %xmm4, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: packuswb %xmm2, %xmm0
; SSE2-NEXT: packuswb %xmm0, %xmm0
; SSE2-NEXT: movq %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i64_8i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0]
; SSSE3-NEXT: pand %xmm4, %xmm3
; SSSE3-NEXT: pand %xmm4, %xmm2
; SSSE3-NEXT: packuswb %xmm3, %xmm2
; SSSE3-NEXT: pand %xmm4, %xmm1
; SSSE3-NEXT: pand %xmm4, %xmm0
; SSSE3-NEXT: packuswb %xmm1, %xmm0
; SSSE3-NEXT: packuswb %xmm2, %xmm0
; SSSE3-NEXT: packuswb %xmm0, %xmm0
; SSSE3-NEXT: movq %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i64_8i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,0,0,0,0,255,0,0,0,0,0,0,0]
; SSE41-NEXT: pand %xmm4, %xmm3
; SSE41-NEXT: pand %xmm4, %xmm2
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: pand %xmm4, %xmm1
; SSE41-NEXT: pand %xmm4, %xmm0
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: packusdw %xmm2, %xmm0
; SSE41-NEXT: packuswb %xmm0, %xmm0
; SSE41-NEXT: movq %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc8i64_8i8:
; AVX1: # %bb.0: # %entry
[X86] combineVectorTruncationWithPACKUS - remove split/concatenation of mask combineVectorTruncationWithPACKUS is currently splitting the upper bit bit masking into 128-bit subregs and then concatenating them back together. This was originally done to avoid regressions that caused existing subregs to be concatenated to the larger type just for the AND masking before being extracted again. This was fixed by @spatel (notably rL303997 and rL347356). This also lets SimplifyDemandedBits do some further improvements before it hits the recursive depth limit. My only annoyance with this is that we were broadcasting some xmm masks but we seem to have lost them by moving to ymm - but that's a known issue as the logic in lowerBuildVectorAsBroadcast isn't great. Differential Revision: https://reviews.llvm.org/D60375#inline-539623 llvm-svn: 358692
2019-04-19 01:23:09 +08:00
; AVX1-NEXT: vmovaps {{.*#+}} ymm2 = [255,255,255,255]
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
; AVX1-NEXT: vpackusdw %xmm3, %xmm1, %xmm1
; AVX1-NEXT: vandps %ymm2, %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpackuswb %xmm0, %xmm0, %xmm0
; AVX1-NEXT: vmovq %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc8i64_8i8:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <u,u,0,8,u,u,u,u,u,u,u,u,u,u,u,u>
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX2-NEXT: vpshufb %xmm3, %xmm1, %xmm1
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3]
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = <0,8,u,u,u,u,u,u,u,u,u,u,u,u,u,u>
; AVX2-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX2-NEXT: vpshufb %xmm3, %xmm0, %xmm0
; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3]
; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2,3]
; AVX2-NEXT: vmovq %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc8i64_8i8:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpmovqb %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = trunc <8 x i64> %a to <8 x i8>
store <8 x i8> %0, <8 x i8>* undef, align 4
ret void
}
define <8 x i16> @trunc8i32_8i16(<8 x i32> %a) {
; SSE2-LABEL: trunc8i32_8i16:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pslld $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: pslld $16, %xmm0
; SSE2-NEXT: psrad $16, %xmm0
; SSE2-NEXT: packssdw %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i32_8i16:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i32_8i16:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSE41-NEXT: pshufb %xmm2, %xmm1
; SSE41-NEXT: pshufb %xmm2, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc8i32_8i16:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc8i32_8i16:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpshufb {{.*#+}} ymm0 = ymm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
; AVX2-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc8i32_8i16:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc8i32_8i16:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc8i32_8i16:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0
; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc8i32_8i16:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = trunc <8 x i32> %a to <8 x i16>
ret <8 x i16> %0
}
define <8 x i16> @trunc8i32_8i16_ashr(<8 x i32> %a) {
; SSE-LABEL: trunc8i32_8i16_ashr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: psrad $16, %xmm1
; SSE-NEXT: psrad $16, %xmm0
; SSE-NEXT: packssdw %xmm1, %xmm0
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc8i32_8i16_ashr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX1-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc8i32_8i16_ashr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc8i32_8i16_ashr:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc8i32_8i16_ashr:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc8i32_8i16_ashr:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0
; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc8i32_8i16_ashr:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX512BWVL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = ashr <8 x i32> %a, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
%1 = trunc <8 x i32> %0 to <8 x i16>
ret <8 x i16> %1
}
define <8 x i16> @trunc8i32_8i16_lshr(<8 x i32> %a) {
; SSE2-LABEL: trunc8i32_8i16_lshr:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm0
; SSE2-NEXT: packssdw %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i32_8i16_lshr:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [2,3,6,7,10,11,14,15,10,11,14,15,14,15,255,255]
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i32_8i16_lshr:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: psrld $16, %xmm1
; SSE41-NEXT: psrld $16, %xmm0
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc8i32_8i16_lshr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vpsrld $16, %xmm1, %xmm1
; AVX1-NEXT: vpsrld $16, %xmm0, %xmm0
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc8i32_8i16_lshr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc8i32_8i16_lshr:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc8i32_8i16_lshr:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc8i32_8i16_lshr:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0
; AVX512BW-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc8i32_8i16_lshr:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX512BWVL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = lshr <8 x i32> %a, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
%1 = trunc <8 x i32> %0 to <8 x i16>
ret <8 x i16> %1
}
define void @trunc8i32_8i8(<8 x i32> %a) {
; SSE2-LABEL: trunc8i32_8i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0]
; SSE2-NEXT: pand %xmm2, %xmm1
; SSE2-NEXT: pand %xmm2, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: packuswb %xmm0, %xmm0
; SSE2-NEXT: movq %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i32_8i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSSE3-NEXT: movq %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i32_8i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
; SSE41-NEXT: pshufb %xmm2, %xmm1
; SSE41-NEXT: pshufb %xmm2, %xmm0
; SSE41-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE41-NEXT: movq %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc8i32_8i8:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
; AVX1-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; AVX1-NEXT: vmovq %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc8i32_8i8:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = <0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u>
; AVX2-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX2-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; AVX2-NEXT: vmovq %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc8i32_8i8:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovdb %zmm0, %xmm0
; AVX512F-NEXT: vmovq %xmm0, (%rax)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc8i32_8i8:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpmovdb %ymm0, (%rax)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc8i32_8i8:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovdb %zmm0, %xmm0
; AVX512BW-NEXT: vmovq %xmm0, (%rax)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc8i32_8i8:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpmovdb %ymm0, (%rax)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = trunc <8 x i32> %a to <8 x i8>
store <8 x i8> %0, <8 x i8>* undef, align 4
ret void
}
define void @trunc16i32_16i16(<16 x i32> %a) {
; SSE2-LABEL: trunc16i32_16i16:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pslld $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: pslld $16, %xmm0
; SSE2-NEXT: psrad $16, %xmm0
; SSE2-NEXT: packssdw %xmm1, %xmm0
; SSE2-NEXT: pslld $16, %xmm3
; SSE2-NEXT: psrad $16, %xmm3
; SSE2-NEXT: pslld $16, %xmm2
; SSE2-NEXT: psrad $16, %xmm2
; SSE2-NEXT: packssdw %xmm3, %xmm2
; SSE2-NEXT: movdqu %xmm2, (%rax)
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc16i32_16i16:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: pslld $16, %xmm1
; SSSE3-NEXT: psrad $16, %xmm1
; SSSE3-NEXT: pslld $16, %xmm0
; SSSE3-NEXT: psrad $16, %xmm0
; SSSE3-NEXT: packssdw %xmm1, %xmm0
; SSSE3-NEXT: pslld $16, %xmm3
; SSSE3-NEXT: psrad $16, %xmm3
; SSSE3-NEXT: pslld $16, %xmm2
; SSSE3-NEXT: psrad $16, %xmm2
; SSSE3-NEXT: packssdw %xmm3, %xmm2
; SSSE3-NEXT: movdqu %xmm2, (%rax)
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc16i32_16i16:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: pxor %xmm4, %xmm4
; SSE41-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0],xmm4[1],xmm1[2],xmm4[3],xmm1[4],xmm4[5],xmm1[6],xmm4[7]
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0],xmm4[1],xmm0[2],xmm4[3],xmm0[4],xmm4[5],xmm0[6],xmm4[7]
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: pblendw {{.*#+}} xmm3 = xmm3[0],xmm4[1],xmm3[2],xmm4[3],xmm3[4],xmm4[5],xmm3[6],xmm4[7]
; SSE41-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0],xmm4[1],xmm2[2],xmm4[3],xmm2[4],xmm4[5],xmm2[6],xmm4[7]
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: movdqu %xmm2, (%rax)
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i16:
; AVX1: # %bb.0: # %entry
[X86] combineVectorTruncationWithPACKUS - remove split/concatenation of mask combineVectorTruncationWithPACKUS is currently splitting the upper bit bit masking into 128-bit subregs and then concatenating them back together. This was originally done to avoid regressions that caused existing subregs to be concatenated to the larger type just for the AND masking before being extracted again. This was fixed by @spatel (notably rL303997 and rL347356). This also lets SimplifyDemandedBits do some further improvements before it hits the recursive depth limit. My only annoyance with this is that we were broadcasting some xmm masks but we seem to have lost them by moving to ymm - but that's a known issue as the logic in lowerBuildVectorAsBroadcast isn't great. Differential Revision: https://reviews.llvm.org/D60375#inline-539623 llvm-svn: 358692
2019-04-19 01:23:09 +08:00
; AVX1-NEXT: vmovaps {{.*#+}} ymm2 = [65535,65535,65535,65535,65535,65535,65535,65535]
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX1-NEXT: vandps %ymm2, %ymm0, %ymm0
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; AVX1-NEXT: vpackusdw %xmm3, %xmm0, %xmm0
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpackusdw %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vmovdqu %xmm1, (%rax)
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i16:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-NEXT: vpshufb %ymm2, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,2,2,3]
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX2-NEXT: vmovdqu %xmm1, (%rax)
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i16:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpmovdw %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = trunc <16 x i32> %a to <16 x i16>
store <16 x i16> %0, <16 x i16>* undef, align 4
ret void
}
define void @trunc16i32_16i16_ashr(<16 x i32> %a) {
; SSE-LABEL: trunc16i32_16i16_ashr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: psrad $16, %xmm3
; SSE-NEXT: psrad $16, %xmm2
; SSE-NEXT: packssdw %xmm3, %xmm2
; SSE-NEXT: psrad $16, %xmm1
; SSE-NEXT: psrad $16, %xmm0
; SSE-NEXT: packssdw %xmm1, %xmm0
; SSE-NEXT: movdqu %xmm2, (%rax)
; SSE-NEXT: movdqu %xmm0, (%rax)
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i16_ashr:
; AVX1: # %bb.0: # %entry
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm2, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm1, %xmm1
; AVX1-NEXT: vpackssdw %xmm2, %xmm1, %xmm1
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm2, %xmm2
; AVX1-NEXT: vpsrad $16, %xmm0, %xmm0
; AVX1-NEXT: vpackssdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm1, (%rax)
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i16_ashr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrad $16, %ymm1, %ymm1
; AVX2-NEXT: vpsrad $16, %ymm0, %ymm0
; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX2-NEXT: vmovdqu %ymm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i16_ashr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsrld $16, %zmm0, %zmm0
; AVX512-NEXT: vpmovdw %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = ashr <16 x i32> %a, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
%1 = trunc <16 x i32> %0 to <16 x i16>
store <16 x i16> %1, <16 x i16>* undef, align 4
ret void
}
define void @trunc16i32_16i16_lshr(<16 x i32> %a) {
; SSE2-LABEL: trunc16i32_16i16_lshr:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm0
; SSE2-NEXT: packssdw %xmm1, %xmm0
; SSE2-NEXT: psrad $16, %xmm3
; SSE2-NEXT: psrad $16, %xmm2
; SSE2-NEXT: packssdw %xmm3, %xmm2
; SSE2-NEXT: movdqu %xmm2, (%rax)
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc16i32_16i16_lshr:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: psrad $16, %xmm1
; SSSE3-NEXT: psrad $16, %xmm0
; SSSE3-NEXT: packssdw %xmm1, %xmm0
; SSSE3-NEXT: psrad $16, %xmm3
; SSSE3-NEXT: psrad $16, %xmm2
; SSSE3-NEXT: packssdw %xmm3, %xmm2
; SSSE3-NEXT: movdqu %xmm2, (%rax)
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc16i32_16i16_lshr:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: psrld $16, %xmm3
; SSE41-NEXT: psrld $16, %xmm2
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: psrld $16, %xmm1
; SSE41-NEXT: psrld $16, %xmm0
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: movdqu %xmm2, (%rax)
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i16_lshr:
; AVX1: # %bb.0: # %entry
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpsrld $16, %xmm2, %xmm2
; AVX1-NEXT: vpsrld $16, %xmm1, %xmm1
; AVX1-NEXT: vpackusdw %xmm2, %xmm1, %xmm1
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpsrld $16, %xmm2, %xmm2
; AVX1-NEXT: vpsrld $16, %xmm0, %xmm0
; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm1, (%rax)
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i16_lshr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrld $16, %ymm1, %ymm1
; AVX2-NEXT: vpsrld $16, %ymm0, %ymm0
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX2-NEXT: vmovdqu %ymm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i16_lshr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsrld $16, %zmm0, %zmm0
; AVX512-NEXT: vpmovdw %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = lshr <16 x i32> %a, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
%1 = trunc <16 x i32> %0 to <16 x i16>
store <16 x i16> %1, <16 x i16>* undef, align 4
ret void
}
define void @trunc16i32_16i8(<16 x i32> %a) {
; SSE2-LABEL: trunc16i32_16i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0]
; SSE2-NEXT: pand %xmm4, %xmm3
; SSE2-NEXT: pand %xmm4, %xmm2
; SSE2-NEXT: packuswb %xmm3, %xmm2
; SSE2-NEXT: pand %xmm4, %xmm1
; SSE2-NEXT: pand %xmm4, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: packuswb %xmm2, %xmm0
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc16i32_16i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0]
; SSSE3-NEXT: pand %xmm4, %xmm3
; SSSE3-NEXT: pand %xmm4, %xmm2
; SSSE3-NEXT: packuswb %xmm3, %xmm2
; SSSE3-NEXT: pand %xmm4, %xmm1
; SSSE3-NEXT: pand %xmm4, %xmm0
; SSSE3-NEXT: packuswb %xmm1, %xmm0
; SSSE3-NEXT: packuswb %xmm2, %xmm0
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc16i32_16i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0]
; SSE41-NEXT: pand %xmm4, %xmm3
; SSE41-NEXT: pand %xmm4, %xmm2
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: pand %xmm4, %xmm1
; SSE41-NEXT: pand %xmm4, %xmm0
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: packuswb %xmm2, %xmm0
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i8:
; AVX1: # %bb.0: # %entry
[X86] combineVectorTruncationWithPACKUS - remove split/concatenation of mask combineVectorTruncationWithPACKUS is currently splitting the upper bit bit masking into 128-bit subregs and then concatenating them back together. This was originally done to avoid regressions that caused existing subregs to be concatenated to the larger type just for the AND masking before being extracted again. This was fixed by @spatel (notably rL303997 and rL347356). This also lets SimplifyDemandedBits do some further improvements before it hits the recursive depth limit. My only annoyance with this is that we were broadcasting some xmm masks but we seem to have lost them by moving to ymm - but that's a known issue as the logic in lowerBuildVectorAsBroadcast isn't great. Differential Revision: https://reviews.llvm.org/D60375#inline-539623 llvm-svn: 358692
2019-04-19 01:23:09 +08:00
; AVX1-NEXT: vmovaps {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255]
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm3
; AVX1-NEXT: vpackusdw %xmm3, %xmm1, %xmm1
; AVX1-NEXT: vandps %ymm2, %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i8:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,2,2,3]
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX2-NEXT: vmovdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
; AVX2-NEXT: vpand %xmm3, %xmm1, %xmm1
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vpshufb %ymm2, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX2-NEXT: vpand %xmm3, %xmm0, %xmm0
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i8:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpmovdb %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = trunc <16 x i32> %a to <16 x i8>
store <16 x i8> %0, <16 x i8>* undef, align 4
ret void
}
define void @trunc16i32_16i8_ashr(<16 x i32> %a) {
; SSE-LABEL: trunc16i32_16i8_ashr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: psrad $24, %xmm1
; SSE-NEXT: psrad $24, %xmm0
; SSE-NEXT: packssdw %xmm1, %xmm0
; SSE-NEXT: psrad $24, %xmm3
; SSE-NEXT: psrad $24, %xmm2
; SSE-NEXT: packssdw %xmm3, %xmm2
; SSE-NEXT: packsswb %xmm2, %xmm0
; SSE-NEXT: movdqu %xmm0, (%rax)
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i8_ashr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm2, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm0, %xmm0
; AVX1-NEXT: vpackssdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm2, %xmm2
; AVX1-NEXT: vpsrad $24, %xmm1, %xmm1
; AVX1-NEXT: vpackssdw %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i8_ashr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrad $24, %ymm1, %ymm1
; AVX2-NEXT: vpsrad $24, %ymm0, %ymm0
; AVX2-NEXT: vpackssdw %ymm1, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i8_ashr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsrld $24, %zmm0, %zmm0
; AVX512-NEXT: vpmovdb %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = ashr <16 x i32> %a, <i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
%1 = trunc <16 x i32> %0 to <16 x i8>
store <16 x i8> %1, <16 x i8>* undef, align 4
ret void
}
define void @trunc16i32_16i8_lshr(<16 x i32> %a) {
; SSE2-LABEL: trunc16i32_16i8_lshr:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: psrld $24, %xmm1
; SSE2-NEXT: psrld $24, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: psrld $24, %xmm3
; SSE2-NEXT: psrld $24, %xmm2
; SSE2-NEXT: packuswb %xmm3, %xmm2
; SSE2-NEXT: packuswb %xmm2, %xmm0
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc16i32_16i8_lshr:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: psrld $24, %xmm1
; SSSE3-NEXT: psrld $24, %xmm0
; SSSE3-NEXT: packuswb %xmm1, %xmm0
; SSSE3-NEXT: psrld $24, %xmm3
; SSSE3-NEXT: psrld $24, %xmm2
; SSSE3-NEXT: packuswb %xmm3, %xmm2
; SSSE3-NEXT: packuswb %xmm2, %xmm0
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc16i32_16i8_lshr:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: psrld $24, %xmm1
; SSE41-NEXT: psrld $24, %xmm0
; SSE41-NEXT: packusdw %xmm1, %xmm0
; SSE41-NEXT: psrld $24, %xmm3
; SSE41-NEXT: psrld $24, %xmm2
; SSE41-NEXT: packusdw %xmm3, %xmm2
; SSE41-NEXT: packuswb %xmm2, %xmm0
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc16i32_16i8_lshr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpsrld $24, %xmm2, %xmm2
; AVX1-NEXT: vpsrld $24, %xmm0, %xmm0
; AVX1-NEXT: vpackusdw %xmm2, %xmm0, %xmm0
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpsrld $24, %xmm2, %xmm2
; AVX1-NEXT: vpsrld $24, %xmm1, %xmm1
; AVX1-NEXT: vpackusdw %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i32_16i8_lshr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrld $24, %ymm1, %ymm1
; AVX2-NEXT: vpsrld $24, %ymm0, %ymm0
; AVX2-NEXT: vpackusdw %ymm1, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512-LABEL: trunc16i32_16i8_lshr:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpsrld $24, %zmm0, %zmm0
; AVX512-NEXT: vpmovdb %zmm0, (%rax)
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
entry:
%0 = lshr <16 x i32> %a, <i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
%1 = trunc <16 x i32> %0 to <16 x i8>
store <16 x i8> %1, <16 x i8>* undef, align 4
ret void
}
;PR25684
define void @trunc16i16_16i8(<16 x i16> %a) {
; SSE2-LABEL: trunc16i16_16i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT: pand %xmm2, %xmm1
; SSE2-NEXT: pand %xmm2, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc16i16_16i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc16i16_16i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSE41-NEXT: pshufb %xmm2, %xmm1
; SSE41-NEXT: pshufb %xmm2, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc16i16_16i8:
; AVX1: # %bb.0: # %entry
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX1-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i16_16i8:
; AVX2: # %bb.0: # %entry
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX2-NEXT: vpand {{.*}}(%rip), %ymm0, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc16i16_16i8:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, (%rax)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc16i16_16i8:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, (%rax)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc16i16_16i8:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT: vmovdqu %xmm0, (%rax)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc16i16_16i8:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpmovwb %ymm0, (%rax)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = trunc <16 x i16> %a to <16 x i8>
store <16 x i8> %0, <16 x i8>* undef, align 4
ret void
}
define void @trunc16i16_16i8_ashr(<16 x i16> %a) {
; SSE-LABEL: trunc16i16_16i8_ashr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: psraw $8, %xmm1
; SSE-NEXT: psraw $8, %xmm0
; SSE-NEXT: packsswb %xmm1, %xmm0
; SSE-NEXT: movdqu %xmm0, (%rax)
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc16i16_16i8_ashr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vpsraw $8, %xmm1, %xmm1
; AVX1-NEXT: vpsraw $8, %xmm0, %xmm0
; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i16_16i8_ashr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsraw $8, %ymm0, %ymm0
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc16i16_16i8_ashr:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpsraw $8, %ymm0, %ymm0
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, (%rax)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc16i16_16i8_ashr:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpsraw $8, %ymm0, %ymm0
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, (%rax)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc16i16_16i8_ashr:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpsraw $8, %ymm0, %ymm0
; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT: vmovdqu %xmm0, (%rax)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc16i16_16i8_ashr:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512BWVL-NEXT: vpmovwb %ymm0, (%rax)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = ashr <16 x i16> %a, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
%1 = trunc <16 x i16> %0 to <16 x i8>
store <16 x i8> %1, <16 x i8>* undef, align 4
ret void
}
define void @trunc16i16_16i8_lshr(<16 x i16> %a) {
; SSE-LABEL: trunc16i16_16i8_lshr:
; SSE: # %bb.0: # %entry
; SSE-NEXT: psrlw $8, %xmm1
; SSE-NEXT: psrlw $8, %xmm0
; SSE-NEXT: packuswb %xmm1, %xmm0
; SSE-NEXT: movdqu %xmm0, (%rax)
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc16i16_16i8_lshr:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
; AVX1-NEXT: vpsrlw $8, %xmm0, %xmm0
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc16i16_16i8_lshr:
; AVX2: # %bb.0: # %entry
; AVX2-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT: vmovdqu %xmm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc16i16_16i8_lshr:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, (%rax)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc16i16_16i8_lshr:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, (%rax)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc16i16_16i8_lshr:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT: vmovdqu %xmm0, (%rax)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc16i16_16i8_lshr:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512BWVL-NEXT: vpmovwb %ymm0, (%rax)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = lshr <16 x i16> %a, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
%1 = trunc <16 x i16> %0 to <16 x i8>
store <16 x i8> %1, <16 x i8>* undef, align 4
ret void
}
define void @trunc32i16_32i8(<32 x i16> %a) {
; SSE2-LABEL: trunc32i16_32i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT: pand %xmm4, %xmm1
; SSE2-NEXT: pand %xmm4, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: pand %xmm4, %xmm3
; SSE2-NEXT: pand %xmm4, %xmm2
; SSE2-NEXT: packuswb %xmm3, %xmm2
; SSE2-NEXT: movdqu %xmm2, (%rax)
; SSE2-NEXT: movdqu %xmm0, (%rax)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc32i16_32i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSSE3-NEXT: pshufb %xmm4, %xmm1
; SSSE3-NEXT: pshufb %xmm4, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: pshufb %xmm4, %xmm3
; SSSE3-NEXT: pshufb %xmm4, %xmm2
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
; SSSE3-NEXT: movdqu %xmm2, (%rax)
; SSSE3-NEXT: movdqu %xmm0, (%rax)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc32i16_32i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSE41-NEXT: pshufb %xmm4, %xmm1
; SSE41-NEXT: pshufb %xmm4, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: pshufb %xmm4, %xmm3
; SSE41-NEXT: pshufb %xmm4, %xmm2
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
; SSE41-NEXT: movdqu %xmm2, (%rax)
; SSE41-NEXT: movdqu %xmm0, (%rax)
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc32i16_32i8:
; AVX1: # %bb.0: # %entry
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX1-NEXT: vmovaps {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX1-NEXT: vandps %ymm2, %ymm0, %ymm0
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; AVX1-NEXT: vpackuswb %xmm3, %xmm0, %xmm0
; AVX1-NEXT: vandps %ymm2, %ymm1, %ymm1
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
; AVX1-NEXT: vmovdqu %xmm1, (%rax)
; AVX1-NEXT: vmovdqu %xmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: trunc32i16_32i8:
; AVX2: # %bb.0: # %entry
[X86] Lower v16i16->v8i16 truncate using an 'and' with 255, an extract_subvector, and a packuswb instruction. Summary: This is an improvement over the two pshufbs and punpcklqdq we'd get otherwise. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D54671 llvm-svn: 347171
2018-11-19 01:59:28 +08:00
; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX2-NEXT: vpand %ymm2, %ymm0, %ymm0
[X86][AVX] Fold concat(packus(),packus()) -> packus(concat(),concat()) (PR34773) Basic "revectorization" combine, we can probably do more opcodes here but it can be a tricky cost-benefit depending on where the subvectors came from - but this case helps shuffle combining. llvm-svn: 360134
2019-05-07 19:17:39 +08:00
; AVX2-NEXT: vpand %ymm2, %ymm1, %ymm1
; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm0
; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; AVX2-NEXT: vpackuswb %ymm0, %ymm2, %ymm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-NEXT: vmovdqu %ymm0, (%rax)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: trunc32i16_32i8:
; AVX512F: # %bb.0: # %entry
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX512F-NEXT: vpmovdb %zmm1, (%rax)
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc32i16_32i8:
; AVX512VL: # %bb.0: # %entry
Revert "[x86] split 256-bit store of concatenated vectors" This reverts commit d5a8637072f4c556b88156bd2f6237a2ead47d31. Most likely suspect for this bot failure: http://lab.llvm.org:8011/builders/clang-cmake-x86_64-avx2-linux/builds/9684 llvm-svn: 361850
2019-05-29 01:37:58 +08:00
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
[x86] split 256-bit store of concatenated vectors This shows up as a side issue to the main problem for the AVX target example from PR37428: https://bugs.llvm.org/show_bug.cgi?id=37428 - https://godbolt.org/z/7tpRa3 But as we can see in the pile of existing test diffs, it's actually a widespread problem that affects any AVX or later target. Apart from a couple of oddballs, I think these are all improvements for the reasons stated in the code comment: we do not want to enable YMM unnecessarily (avoid vzeroupper and frequency throttling) and some cores split 256-bit stores anyway. We could say that MergeConsecutiveStores() is going overboard on some of these examples, but that won't solve the problem completely. But that is a reason I'm proposing this as a lowering rather than a combine: we will infinite loop fighting the merge code if we try this earlier. Differential Revision: https://reviews.llvm.org/D62498 llvm-svn: 362524
2019-06-05 00:40:04 +08:00
; AVX512VL-NEXT: vpmovdb %zmm1, (%rax)
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, (%rax)
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc32i16_32i8:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vpmovwb %zmm0, (%rax)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc32i16_32i8:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpmovwb %zmm0, (%rax)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
entry:
%0 = trunc <32 x i16> %a to <32 x i8>
store <32 x i8> %0, <32 x i8>* undef, align 4
ret void
}
define <8 x i32> @trunc2x4i64_8i32(<4 x i64> %a, <4 x i64> %b) {
; SSE-LABEL: trunc2x4i64_8i32:
; SSE: # %bb.0: # %entry
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
; SSE-NEXT: shufps {{.*#+}} xmm2 = xmm2[0,2],xmm3[0,2]
; SSE-NEXT: movaps %xmm2, %xmm1
; SSE-NEXT: retq
;
; AVX1-LABEL: trunc2x4i64_8i32:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc2x4i64_8i32:
; AVX2-SLOW: # %bb.0: # %entry
[x86] split more v8f32/v8i32 shuffles in lowering Similar to D57867 - this is a small patch with lots of test diffs. With half-vector-width narrowing potential, using an extract + 128-bit vshufps is a win because it replaces a 256-bit shuffle with a 128-bit shufle. This seems like it should be a win even for targets with 'fast-variable-shuffle', but we are intentionally deferring that to an independent change to make sure that is true. Differential Revision: https://reviews.llvm.org/D58181 llvm-svn: 354279
2019-02-19 00:46:12 +08:00
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; AVX2-SLOW-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vshufps {{.*#+}} xmm1 = xmm1[0,2],xmm2[0,2]
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX2-SLOW-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc2x4i64_8i32:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = [0,2,4,6,4,6,6,7]
; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT: vpermps %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: trunc2x4i64_8i32:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovqd %zmm0, %ymm0
; AVX512F-NEXT: vpmovqd %zmm1, %ymm1
; AVX512F-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc2x4i64_8i32:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vpmovqd %ymm0, %xmm0
; AVX512VL-NEXT: vpmovqd %ymm1, %xmm1
; AVX512VL-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc2x4i64_8i32:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovqd %zmm0, %ymm0
; AVX512BW-NEXT: vpmovqd %zmm1, %ymm1
; AVX512BW-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc2x4i64_8i32:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vpmovqd %ymm0, %xmm0
; AVX512BWVL-NEXT: vpmovqd %ymm1, %xmm1
; AVX512BWVL-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
; AVX512BWVL-NEXT: retq
entry:
%0 = trunc <4 x i64> %a to <4 x i32>
%1 = trunc <4 x i64> %b to <4 x i32>
%2 = shufflevector <4 x i32> %0, <4 x i32> %1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i32> %2
}
define <8 x i16> @trunc2x4i64_8i16(<4 x i64> %a, <4 x i64> %b) {
; SSE2-LABEL: trunc2x4i64_8i16:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
; SSE2-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE2-NEXT: movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc2x4i64_8i16:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm4 = xmm0[0,2,2,3,4,5,6,7]
; SSSE3-NEXT: punpckldq {{.*#+}} xmm4 = xmm4[0],xmm1[0],xmm4[1],xmm1[1]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm0[0,1,0,2,4,5,6,7]
; SSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[0,2,2,3]
; SSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,1,0,2,4,5,6,7]
; SSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSSE3-NEXT: movsd {{.*#+}} xmm0 = xmm4[0],xmm0[1]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc2x4i64_8i16:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
; SSE41-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[0,1,0,2,4,5,6,7]
; SSE41-NEXT: pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; SSE41-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
; SSE41-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1]
; SSE41-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE41-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE41-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
; SSE41-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm2[4,5,6,7]
; SSE41-NEXT: retq
;
; AVX1-LABEL: trunc2x4i64_8i16:
; AVX1: # %bb.0: # %entry
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; AVX1-NEXT: vpshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
; AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; AVX1-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
; AVX1-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; AVX1-NEXT: vpshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
; AVX1-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; AVX1-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
; AVX1-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1,2,3],xmm1[4,5,6,7]
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: trunc2x4i64_8i16:
; AVX2-SLOW: # %bb.0: # %entry
; AVX2-SLOW-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm2 = xmm2[0,1,0,2,4,5,6,7]
; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,1,0,2,4,5,6,7]
; AVX2-SLOW-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; AVX2-SLOW-NEXT: vextracti128 $1, %ymm0, %xmm2
; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm2 = xmm2[0,2,2,3,4,5,6,7]
; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
; AVX2-SLOW-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; AVX2-SLOW-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
; AVX2-SLOW-NEXT: vzeroupper
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: trunc2x4i64_8i16:
; AVX2-FAST: # %bb.0: # %entry
; AVX2-FAST-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2-FAST-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,2,3,0,1,8,9,8,9,10,11,12,13,14,15]
; AVX2-FAST-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX2-FAST-NEXT: vpshufb %xmm3, %xmm1, %xmm1
; AVX2-FAST-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; AVX2-FAST-NEXT: vextracti128 $1, %ymm0, %xmm2
; AVX2-FAST-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,8,9,8,9,10,11,8,9,10,11,12,13,14,15]
; AVX2-FAST-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX2-FAST-NEXT: vpshufb %xmm3, %xmm0, %xmm0
; AVX2-FAST-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; AVX2-FAST-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3]
; AVX2-FAST-NEXT: vzeroupper
; AVX2-FAST-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512F-LABEL: trunc2x4i64_8i16:
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512F: # %bb.0: # %entry
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512F-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovqw %zmm0, %xmm0
; AVX512F-NEXT: vpmovqw %zmm1, %xmm1
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512F-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512VL-LABEL: trunc2x4i64_8i16:
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512VL: # %bb.0: # %entry
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512VL-NEXT: vpmovqw %ymm0, %xmm0
; AVX512VL-NEXT: vpmovqw %ymm1, %xmm1
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512VL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512BW-LABEL: trunc2x4i64_8i16:
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512BW: # %bb.0: # %entry
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512BW-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovqw %zmm0, %xmm0
; AVX512BW-NEXT: vpmovqw %zmm1, %xmm1
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512BW-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512BWVL-LABEL: trunc2x4i64_8i16:
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512BWVL: # %bb.0: # %entry
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512BWVL-NEXT: vpmovqw %ymm0, %xmm0
; AVX512BWVL-NEXT: vpmovqw %ymm1, %xmm1
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX512BWVL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
entry:
%0 = trunc <4 x i64> %a to <4 x i16>
%1 = trunc <4 x i64> %b to <4 x i16>
%2 = shufflevector <4 x i16> %0, <4 x i16> %1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i16> %2
}
define <4 x i32> @trunc2x2i64_4i32(<2 x i64> %a, <2 x i64> %b) {
; SSE-LABEL: trunc2x2i64_4i32:
; SSE: # %bb.0: # %entry
; SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
; SSE-NEXT: retq
;
; AVX-LABEL: trunc2x2i64_4i32:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
; AVX-NEXT: retq
;
; AVX512-LABEL: trunc2x2i64_4i32:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,2]
; AVX512-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <2 x i64> %a to <2 x i32>
%1 = trunc <2 x i64> %b to <2 x i32>
%2 = shufflevector <2 x i32> %0, <2 x i32> %1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
ret <4 x i32> %2
}
define i64 @trunc2i64_i64(<2 x i64> %inval) {
; SSE-LABEL: trunc2i64_i64:
; SSE: # %bb.0: # %entry
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE-NEXT: movq %xmm0, %rax
; SSE-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX-LABEL: trunc2i64_i64:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; AVX-NEXT: vmovq %xmm0, %rax
; AVX-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512-LABEL: trunc2i64_i64:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; AVX512-NEXT: vmovq %xmm0, %rax
; AVX512-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <2 x i64> %inval to <2 x i32>
%1 = bitcast <2 x i32> %0 to i64
ret i64 %1
}
define <8 x i16> @trunc2x4i32_8i16(<4 x i32> %a, <4 x i32> %b) {
; SSE2-LABEL: trunc2x4i32_8i16:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
; SSE2-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc2x4i32_8i16:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc2x4i32_8i16:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSE41-NEXT: pshufb %xmm2, %xmm1
; SSE41-NEXT: pshufb %xmm2, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX-LABEL: trunc2x4i32_8i16:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
;
[X86][AVX] Combine non-lane crossing binary shuffles using X86ISD::VPERMV3 Some of the combines might be further improved if we lower more shuffles with X86ISD::VPERMV3 directly, instead of waiting to combine the results. llvm-svn: 359400
2019-04-28 22:31:01 +08:00
; AVX512F-LABEL: trunc2x4i32_8i16:
; AVX512F: # %bb.0: # %entry
; AVX512F-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX512F-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX512F-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX512F-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: trunc2x4i32_8i16:
; AVX512VL: # %bb.0: # %entry
; AVX512VL-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX512VL-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX512VL-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX512VL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: trunc2x4i32_8i16:
; AVX512BW: # %bb.0: # %entry
; AVX512BW-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX512BW-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX512BW-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX512BW-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: trunc2x4i32_8i16:
; AVX512BWVL: # %bb.0: # %entry
; AVX512BWVL-NEXT: vmovdqa {{.*#+}} xmm2 = [0,2,4,6,8,10,12,14]
; AVX512BWVL-NEXT: vpermt2w %xmm1, %xmm2, %xmm0
; AVX512BWVL-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <4 x i32> %a to <4 x i16>
%1 = trunc <4 x i32> %b to <4 x i16>
%2 = shufflevector <4 x i16> %0, <4 x i16> %1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i16> %2
}
; PR15524 http://llvm.org/bugs/show_bug.cgi?id=15524
define i64 @trunc4i32_i64(<4 x i32> %inval) {
; SSE2-LABEL: trunc4i32_i64:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
; SSE2-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT: movq %xmm0, %rax
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc4i32_i64:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSSE3-NEXT: movq %xmm0, %rax
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc4i32_i64:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSE41-NEXT: movq %xmm0, %rax
; SSE41-NEXT: retq
;
; AVX-LABEL: trunc4i32_i64:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX-NEXT: vmovq %xmm0, %rax
; AVX-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512-LABEL: trunc4i32_i64:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX512-NEXT: vmovq %xmm0, %rax
; AVX512-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <4 x i32> %inval to <4 x i16>
%1 = bitcast <4 x i16> %0 to i64
ret i64 %1
}
define <16 x i8> @trunc2x8i16_16i8(<8 x i16> %a, <8 x i16> %b) {
; SSE2-LABEL: trunc2x8i16_16i8:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT: pand %xmm2, %xmm1
; SSE2-NEXT: pand %xmm2, %xmm0
; SSE2-NEXT: packuswb %xmm1, %xmm0
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc2x8i16_16i8:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: movdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSSE3-NEXT: pshufb %xmm2, %xmm1
; SSSE3-NEXT: pshufb %xmm2, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc2x8i16_16i8:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: movdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; SSE41-NEXT: pshufb %xmm2, %xmm1
; SSE41-NEXT: pshufb %xmm2, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: retq
;
; AVX-LABEL: trunc2x8i16_16i8:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; AVX-NEXT: vpshufb %xmm2, %xmm1, %xmm1
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX-NEXT: vpshufb %xmm2, %xmm0, %xmm0
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512-LABEL: trunc2x8i16_16i8:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vmovdqa {{.*#+}} xmm2 = <0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u>
; AVX512-NEXT: vpshufb %xmm2, %xmm1, %xmm1
; AVX512-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX512-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <8 x i16> %a to <8 x i8>
%1 = trunc <8 x i16> %b to <8 x i8>
%2 = shufflevector <8 x i8> %0, <8 x i8> %1, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
ret <16 x i8> %2
}
; PR15524 http://llvm.org/bugs/show_bug.cgi?id=15524
define i64 @trunc8i16_i64(<8 x i16> %inval) {
; SSE2-LABEL: trunc8i16_i64:
; SSE2: # %bb.0: # %entry
; SSE2-NEXT: pand {{.*}}(%rip), %xmm0
; SSE2-NEXT: packuswb %xmm0, %xmm0
; SSE2-NEXT: movq %xmm0, %rax
; SSE2-NEXT: retq
;
; SSSE3-LABEL: trunc8i16_i64:
; SSSE3: # %bb.0: # %entry
; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; SSSE3-NEXT: movq %xmm0, %rax
; SSSE3-NEXT: retq
;
; SSE41-LABEL: trunc8i16_i64:
; SSE41: # %bb.0: # %entry
; SSE41-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; SSE41-NEXT: movq %xmm0, %rax
; SSE41-NEXT: retq
;
; AVX-LABEL: trunc8i16_i64:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; AVX-NEXT: vmovq %xmm0, %rax
; AVX-NEXT: retq
;
[X86] Add some custom type legalization rules for truncate with -x86-experimental-vector-widening-legalization. This avoids some nasty shuffles when we have avx512. It will also prevent using zmm truncate instructions when a ymm instruction that zeroes part of an xmm register will do. Also avoid using avx512 truncate instructions when the input is 128 bits or less. These instructions are 2 uops on skx so we can probably find a better single uop shuffle like pshufb. llvm-svn: 346936
2018-11-15 16:23:40 +08:00
; AVX512-LABEL: trunc8i16_i64:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,2,4,6,8,10,12,14,u,u,u,u,u,u,u,u]
; AVX512-NEXT: vmovq %xmm0, %rax
; AVX512-NEXT: retq
[X86] Add more tests for -x86-experimental-vector-widening-legalization I'm looking into whether we can make this the default legalization strategy. Adding these tests to help cover the changes that will be necessary. This patch adds copies of some tests with the command line switch enabled. By making copies its easier to compare the two legalization strategies. I've also removed RUN lines from some of these tests that already had -x86-experimental-vector-widening-legalization llvm-svn: 346745
2018-11-13 15:47:52 +08:00
entry:
%0 = trunc <8 x i16> %inval to <8 x i8>
%1 = bitcast <8 x i8> %0 to i64
ret i64 %1
}
define <16 x i8> @trunc16i64_16i8_const() {
; SSE-LABEL: trunc16i64_16i8_const:
; SSE: # %bb.0: # %entry
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: trunc16i64_16i8_const:
; AVX: # %bb.0: # %entry
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: retq
;
; AVX512-LABEL: trunc16i64_16i8_const:
; AVX512: # %bb.0: # %entry
; AVX512-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX512-NEXT: retq
entry:
%0 = trunc <16 x i64> zeroinitializer to <16 x i8>
%1 = shufflevector <16 x i8> %0, <16 x i8> %0, <16 x i32> <i32 28, i32 30, i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 undef, i32 14, i32 16, i32 18, i32 20, i32 22, i32 24, i32 26>
ret <16 x i8> %1
}
define <8 x i16> @PR32160(<8 x i32> %x) {
; SSE-LABEL: PR32160:
; SSE: # %bb.0:
; SSE-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,6,7]
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,2,2,2]
; SSE-NEXT: retq
;
; AVX1-LABEL: PR32160:
; AVX1: # %bb.0:
; AVX1-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[8,9,8,9,8,9,8,9,8,9,8,9,8,9,8,9]
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-SLOW-LABEL: PR32160:
; AVX2-SLOW: # %bb.0:
; AVX2-SLOW-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4,4,5,5,6,6,7,7]
; AVX2-SLOW-NEXT: vpbroadcastd %xmm0, %xmm0
; AVX2-SLOW-NEXT: vzeroupper
; AVX2-SLOW-NEXT: retq
;
; AVX2-FAST-LABEL: PR32160:
; AVX2-FAST: # %bb.0:
; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[8,9,8,9,8,9,8,9,8,9,8,9,8,9,8,9]
; AVX2-FAST-NEXT: vzeroupper
; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: PR32160:
; AVX512F: # %bb.0:
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
; AVX512F-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[2,2,2,3,4,5,6,7]
; AVX512F-NEXT: vpbroadcastd %xmm0, %xmm0
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: PR32160:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,5,4,5,4,5,4,5,4,5,4,5,4,5,4,5]
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: PR32160:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0
; AVX512BW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,5,4,5,4,5,4,5,4,5,4,5,4,5,4,5]
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: PR32160:
; AVX512BWVL: # %bb.0:
; AVX512BWVL-NEXT: vpmovdw %ymm0, %xmm0
; AVX512BWVL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[4,5,4,5,4,5,4,5,4,5,4,5,4,5,4,5]
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
%shuf = trunc <8 x i32> %x to <8 x i16>
%trunc = shufflevector <8 x i16> %shuf, <8 x i16> undef, <8 x i32> <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
ret <8 x i16> %trunc
}
define void @PR34773(i16* %a0, i8* %a1) {
; SSE-LABEL: PR34773:
; SSE: # %bb.0:
; SSE-NEXT: movdqu (%rdi), %xmm0
; SSE-NEXT: movdqu 16(%rdi), %xmm1
; SSE-NEXT: movdqu 32(%rdi), %xmm2
; SSE-NEXT: movdqu 48(%rdi), %xmm3
; SSE-NEXT: psrlw $8, %xmm1
; SSE-NEXT: psrlw $8, %xmm0
; SSE-NEXT: packuswb %xmm1, %xmm0
; SSE-NEXT: psrlw $8, %xmm3
; SSE-NEXT: psrlw $8, %xmm2
; SSE-NEXT: packuswb %xmm3, %xmm2
; SSE-NEXT: movdqu %xmm0, (%rsi)
; SSE-NEXT: movdqu %xmm2, 16(%rsi)
; SSE-NEXT: retq
;
; AVX1-LABEL: PR34773:
; AVX1: # %bb.0:
; AVX1-NEXT: vmovdqu (%rdi), %xmm0
; AVX1-NEXT: vmovdqu 16(%rdi), %xmm1
; AVX1-NEXT: vmovdqu 32(%rdi), %xmm2
; AVX1-NEXT: vmovdqu 48(%rdi), %xmm3
; AVX1-NEXT: vpsrlw $8, %xmm1, %xmm1
; AVX1-NEXT: vpsrlw $8, %xmm0, %xmm0
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
; AVX1-NEXT: vpsrlw $8, %xmm3, %xmm1
; AVX1-NEXT: vpsrlw $8, %xmm2, %xmm2
; AVX1-NEXT: vpackuswb %xmm1, %xmm2, %xmm1
; AVX1-NEXT: vmovdqu %xmm0, (%rsi)
; AVX1-NEXT: vmovdqu %xmm1, 16(%rsi)
; AVX1-NEXT: retq
;
; AVX2-LABEL: PR34773:
; AVX2: # %bb.0:
; AVX2-NEXT: vmovdqu (%rdi), %ymm0
; AVX2-NEXT: vmovdqu 32(%rdi), %ymm1
; AVX2-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX2-NEXT: vpsrlw $8, %ymm1, %ymm1
; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm2
; AVX2-NEXT: vpackuswb %xmm2, %xmm0, %xmm0
; AVX2-NEXT: vextracti128 $1, %ymm1, %xmm2
; AVX2-NEXT: vpackuswb %xmm2, %xmm1, %xmm1
; AVX2-NEXT: vmovdqu %xmm0, (%rsi)
; AVX2-NEXT: vmovdqu %xmm1, 16(%rsi)
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: PR34773:
; AVX512F: # %bb.0:
; AVX512F-NEXT: vmovdqu (%rdi), %ymm0
; AVX512F-NEXT: vmovdqu 32(%rdi), %ymm1
; AVX512F-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512F-NEXT: vpsrlw $8, %ymm1, %ymm1
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, (%rsi)
; AVX512F-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
; AVX512F-NEXT: vpmovdb %zmm0, 16(%rsi)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: PR34773:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vmovdqu (%rdi), %ymm0
; AVX512VL-NEXT: vmovdqu 32(%rdi), %ymm1
; AVX512VL-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512VL-NEXT: vpsrlw $8, %ymm1, %ymm1
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, (%rsi)
; AVX512VL-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero
; AVX512VL-NEXT: vpmovdb %zmm0, 16(%rsi)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: PR34773:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vmovdqu (%rdi), %ymm0
; AVX512BW-NEXT: vmovdqu 32(%rdi), %ymm1
; AVX512BW-NEXT: vpsrlw $8, %ymm0, %ymm0
; AVX512BW-NEXT: vpsrlw $8, %ymm1, %ymm1
; AVX512BW-NEXT: vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT: vpmovwb %zmm1, %ymm1
; AVX512BW-NEXT: vmovdqu %xmm0, (%rsi)
; AVX512BW-NEXT: vmovdqu %xmm1, 16(%rsi)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: PR34773:
; AVX512BWVL: # %bb.0:
; AVX512BWVL-NEXT: vpsrlw $8, (%rdi), %ymm0
; AVX512BWVL-NEXT: vpsrlw $8, 32(%rdi), %ymm1
; AVX512BWVL-NEXT: vpmovwb %ymm0, (%rsi)
; AVX512BWVL-NEXT: vpmovwb %ymm1, 16(%rsi)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
%1 = getelementptr i16, i16* %a0, i64 16
%2 = getelementptr i8, i8* %a1, i64 16
%3 = bitcast i16* %a0 to <16 x i16>*
%4 = bitcast i16* %1 to <16 x i16>*
%5 = bitcast i8* %a1 to <16 x i8>*
%6 = bitcast i8* %2 to <16 x i8>*
%7 = load <16 x i16>, <16 x i16>* %3, align 2
%8 = load <16 x i16>, <16 x i16>* %4, align 2
%9 = lshr <16 x i16> %7, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
%10 = lshr <16 x i16> %8, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
%11 = trunc <16 x i16> %9 to <16 x i8>
%12 = trunc <16 x i16> %10 to <16 x i8>
store <16 x i8> %11, <16 x i8>* %5, align 1
store <16 x i8> %12, <16 x i8>* %6, align 1
ret void
}
[x86] add test for store merging/splitting; NFC This is a reduction of a test that would infinite loop with D62498. llvm-svn: 362502
2019-06-04 22:40:37 +08:00
; Store merging must not infinitely fight store splitting.
define void @store_merge_split(<8 x i32> %w1, <8 x i32> %w2, i64 %idx, <8 x i16>* %p) align 2 {
; SSE2-LABEL: store_merge_split:
; SSE2: # %bb.0:
; SSE2-NEXT: pslld $16, %xmm1
; SSE2-NEXT: psrad $16, %xmm1
; SSE2-NEXT: pslld $16, %xmm0
; SSE2-NEXT: psrad $16, %xmm0
; SSE2-NEXT: packssdw %xmm1, %xmm0
; SSE2-NEXT: pslld $16, %xmm3
; SSE2-NEXT: psrad $16, %xmm3
; SSE2-NEXT: pslld $16, %xmm2
; SSE2-NEXT: psrad $16, %xmm2
; SSE2-NEXT: packssdw %xmm3, %xmm2
; SSE2-NEXT: shlq $4, %rdi
; SSE2-NEXT: movdqu %xmm0, (%rsi,%rdi)
; SSE2-NEXT: movdqu %xmm2, 16(%rsi,%rdi)
; SSE2-NEXT: retq
;
; SSSE3-LABEL: store_merge_split:
; SSSE3: # %bb.0:
; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSSE3-NEXT: pshufb %xmm4, %xmm1
; SSSE3-NEXT: pshufb %xmm4, %xmm0
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSSE3-NEXT: pshufb %xmm4, %xmm3
; SSSE3-NEXT: pshufb %xmm4, %xmm2
; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
; SSSE3-NEXT: shlq $4, %rdi
; SSSE3-NEXT: movdqu %xmm0, (%rsi,%rdi)
; SSSE3-NEXT: movdqu %xmm2, 16(%rsi,%rdi)
; SSSE3-NEXT: retq
;
; SSE41-LABEL: store_merge_split:
; SSE41: # %bb.0:
; SSE41-NEXT: movdqa {{.*#+}} xmm4 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; SSE41-NEXT: pshufb %xmm4, %xmm1
; SSE41-NEXT: pshufb %xmm4, %xmm0
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; SSE41-NEXT: pshufb %xmm4, %xmm3
; SSE41-NEXT: pshufb %xmm4, %xmm2
; SSE41-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm3[0]
; SSE41-NEXT: shlq $4, %rdi
; SSE41-NEXT: movdqu %xmm0, (%rsi,%rdi)
; SSE41-NEXT: movdqu %xmm2, 16(%rsi,%rdi)
; SSE41-NEXT: retq
;
; AVX1-LABEL: store_merge_split:
; AVX1: # %bb.0:
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX1-NEXT: vpshufb %xmm3, %xmm0, %xmm0
; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm2[0]
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; AVX1-NEXT: vpshufb %xmm3, %xmm2, %xmm2
; AVX1-NEXT: vpshufb %xmm3, %xmm1, %xmm1
; AVX1-NEXT: vpunpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm2[0]
; AVX1-NEXT: shlq $4, %rdi
; AVX1-NEXT: vmovdqu %xmm0, (%rsi,%rdi)
; AVX1-NEXT: vmovdqu %xmm1, 16(%rsi,%rdi)
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: store_merge_split:
; AVX2: # %bb.0:
; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15,16,17,20,21,24,25,28,29,24,25,28,29,28,29,30,31]
; AVX2-NEXT: vpshufb %ymm2, %ymm0, %ymm0
; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
; AVX2-NEXT: vpshufb %ymm2, %ymm1, %ymm1
; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,2,2,3]
; AVX2-NEXT: shlq $4, %rdi
[SelectionDAG][x86] limit post-legalization store merging by type The proposal in D62498 showed that x86 would benefit from vector store splitting, but that may conflict with the generic DAG combiner's store merging transforms. Add memory type to the existing TLI hook that enables the merging transforms, so we can limit those changes to scalars only for x86. llvm-svn: 362507
2019-06-04 23:15:59 +08:00
; AVX2-NEXT: vmovdqu %xmm0, (%rsi,%rdi)
; AVX2-NEXT: vmovdqu %xmm1, 16(%rsi,%rdi)
[x86] add test for store merging/splitting; NFC This is a reduction of a test that would infinite loop with D62498. llvm-svn: 362502
2019-06-04 22:40:37 +08:00
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512F-LABEL: store_merge_split:
; AVX512F: # %bb.0:
; AVX512F-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512F-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
; AVX512F-NEXT: vpmovdw %zmm1, %ymm1
; AVX512F-NEXT: shlq $4, %rdi
; AVX512F-NEXT: vmovdqu %xmm0, (%rsi,%rdi)
; AVX512F-NEXT: vmovdqu %xmm1, 16(%rsi,%rdi)
; AVX512F-NEXT: vzeroupper
; AVX512F-NEXT: retq
;
; AVX512VL-LABEL: store_merge_split:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: shlq $4, %rdi
; AVX512VL-NEXT: vpmovdw %ymm0, (%rsi,%rdi)
; AVX512VL-NEXT: vpmovdw %ymm1, 16(%rsi,%rdi)
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: store_merge_split:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512BW-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512BW-NEXT: vpmovdw %zmm0, %ymm0
; AVX512BW-NEXT: vpmovdw %zmm1, %ymm1
; AVX512BW-NEXT: shlq $4, %rdi
; AVX512BW-NEXT: vmovdqu %xmm0, (%rsi,%rdi)
; AVX512BW-NEXT: vmovdqu %xmm1, 16(%rsi,%rdi)
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
;
; AVX512BWVL-LABEL: store_merge_split:
; AVX512BWVL: # %bb.0:
; AVX512BWVL-NEXT: shlq $4, %rdi
; AVX512BWVL-NEXT: vpmovdw %ymm0, (%rsi,%rdi)
; AVX512BWVL-NEXT: vpmovdw %ymm1, 16(%rsi,%rdi)
; AVX512BWVL-NEXT: vzeroupper
; AVX512BWVL-NEXT: retq
%t1 = trunc <8 x i32> %w1 to <8 x i16>
%t2 = trunc <8 x i32> %w2 to <8 x i16>
%g1 = getelementptr inbounds <8 x i16>, <8 x i16>* %p, i64 %idx
%g2 = getelementptr inbounds <8 x i16>, <8 x i16>* %g1, i64 1
store <8 x i16> %t1, <8 x i16>* %g1, align 2
store <8 x i16> %t2, <8 x i16>* %g2, align 2
ret void
}