From 2b55e751ce61d7690706748926327166fa2cd59d Mon Sep 17 00:00:00 2001 From: Simon Pilgrim Date: Sat, 2 Jun 2018 21:35:48 +0000 Subject: [PATCH] [X86][SSE] Cleanup AVX1 intrinsics tests Ensure we cover 32/64-bit targets for SSE/AVX/AVX512 cases as necessary, strip -mcpu usage. llvm-svn: 333834 --- .../CodeGen/X86/avx-intrinsics-x86-upgrade.ll | 843 ++++++++++++------ llvm/test/CodeGen/X86/avx-intrinsics-x86.ll | 8 +- .../test/CodeGen/X86/avx-intrinsics-x86_64.ll | 36 +- 3 files changed, 614 insertions(+), 273 deletions(-) diff --git a/llvm/test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll b/llvm/test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll index f557d7ab0da1..21612506f851 100644 --- a/llvm/test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll +++ b/llvm/test/CodeGen/X86/avx-intrinsics-x86-upgrade.ll @@ -1,34 +1,51 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=CHECK --check-prefix=X86 -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=CHECK --check-prefix=X64 +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86,AVX,X86-AVX +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86,AVX512VL,X86-AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64,AVX,X64-AVX +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64,AVX512VL,X64-AVX512VL ; We don't check any vinsertf128 variant with immediate 0 because that's just a blend. define <4 x double> @test_x86_avx_vinsertf128_pd_256_1(<4 x double> %a0, <2 x double> %a1) { -; CHECK-LABEL: test_x86_avx_vinsertf128_pd_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vinsertf128_pd_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vinsertf128_pd_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.vinsertf128.pd.256(<4 x double> %a0, <2 x double> %a1, i8 1) ret <4 x double> %res } declare <4 x double> @llvm.x86.avx.vinsertf128.pd.256(<4 x double>, <2 x double>, i8) nounwind readnone define <8 x float> @test_x86_avx_vinsertf128_ps_256_1(<8 x float> %a0, <4 x float> %a1) { -; CHECK-LABEL: test_x86_avx_vinsertf128_ps_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vinsertf128_ps_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vinsertf128_ps_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float> %a0, <4 x float> %a1, i8 1) ret <8 x float> %res } declare <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float>, <4 x float>, i8) nounwind readnone define <8 x i32> @test_x86_avx_vinsertf128_si_256_1(<8 x i32> %a0, <4 x i32> %a1) { -; CHECK-LABEL: test_x86_avx_vinsertf128_si_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vinsertf128_si_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vinsertf128_si_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i32> @llvm.x86.avx.vinsertf128.si.256(<8 x i32> %a0, <4 x i32> %a1, i8 1) ret <8 x i32> %res } @@ -40,8 +57,9 @@ define <8 x i32> @test_x86_avx_vinsertf128_si_256_2(<8 x i32> %a0, <4 x i32> %a1 ; CHECK-LABEL: test_x86_avx_vinsertf128_si_256_2: ; CHECK: # %bb.0: ; CHECK-NEXT: # kill: def $xmm1 killed $xmm1 def $ymm1 -; CHECK-NEXT: vblendps {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7] -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vblendps $240, %ymm0, %ymm1, %ymm0 # encoding: [0xc4,0xe3,0x75,0x0c,0xc0,0xf0] +; CHECK-NEXT: # ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i32> @llvm.x86.avx.vinsertf128.si.256(<8 x i32> %a0, <4 x i32> %a1, i8 2) ret <8 x i32> %res } @@ -50,33 +68,51 @@ declare <8 x i32> @llvm.x86.avx.vinsertf128.si.256(<8 x i32>, <4 x i32>, i8) nou ; We don't check any vextractf128 variant with immediate 0 because that's just a move. define <2 x double> @test_x86_avx_vextractf128_pd_256_1(<4 x double> %a0) { -; CHECK-LABEL: test_x86_avx_vextractf128_pd_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vextractf128 $1, %ymm0, %xmm0 -; CHECK-NEXT: vzeroupper -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vextractf128_pd_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vextractf128 $1, %ymm0, %xmm0 # encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vextractf128_pd_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vextractf128 $1, %ymm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double> %a0, i8 1) ret <2 x double> %res } declare <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double>, i8) nounwind readnone define <4 x float> @test_x86_avx_vextractf128_ps_256_1(<8 x float> %a0) { -; CHECK-LABEL: test_x86_avx_vextractf128_ps_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vextractf128 $1, %ymm0, %xmm0 -; CHECK-NEXT: vzeroupper -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vextractf128_ps_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vextractf128 $1, %ymm0, %xmm0 # encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vextractf128_ps_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vextractf128 $1, %ymm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x float> @llvm.x86.avx.vextractf128.ps.256(<8 x float> %a0, i8 1) ret <4 x float> %res } declare <4 x float> @llvm.x86.avx.vextractf128.ps.256(<8 x float>, i8) nounwind readnone define <4 x i32> @test_x86_avx_vextractf128_si_256_1(<8 x i32> %a0) { -; CHECK-LABEL: test_x86_avx_vextractf128_si_256_1: -; CHECK: # %bb.0: -; CHECK-NEXT: vextractf128 $1, %ymm0, %xmm0 -; CHECK-NEXT: vzeroupper -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vextractf128_si_256_1: +; AVX: # %bb.0: +; AVX-NEXT: vextractf128 $1, %ymm0, %xmm0 # encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vextractf128_si_256_1: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vextractf128 $1, %ymm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x19,0xc0,0x01] +; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x i32> @llvm.x86.avx.vextractf128.si.256(<8 x i32> %a0, i8 1) ret <4 x i32> %res } @@ -89,24 +125,39 @@ define <2 x double> @test_x86_avx_extractf128_pd_256_2(<4 x double> %a0) { ; CHECK-LABEL: test_x86_avx_extractf128_pd_256_2: ; CHECK: # %bb.0: ; CHECK-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 -; CHECK-NEXT: vzeroupper -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.avx.vextractf128.pd.256(<4 x double> %a0, i8 2) ret <2 x double> %res } define <4 x double> @test_x86_avx_vbroadcastf128_pd_256(i8* %a0) { -; X86-LABEL: test_x86_avx_vbroadcastf128_pd_256: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1] -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_avx_vbroadcastf128_pd_256: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vbroadcastf128 (%eax), %ymm0 # encoding: [0xc4,0xe2,0x7d,0x1a,0x00] +; X86-AVX-NEXT: # ymm0 = mem[0,1,0,1] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_avx_vbroadcastf128_pd_256: -; X64: # %bb.0: -; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1] -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_avx_vbroadcastf128_pd_256: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vbroadcastf128 (%eax), %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x1a,0x00] +; X86-AVX512VL-NEXT: # ymm0 = mem[0,1,0,1] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_avx_vbroadcastf128_pd_256: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vbroadcastf128 (%rdi), %ymm0 # encoding: [0xc4,0xe2,0x7d,0x1a,0x07] +; X64-AVX-NEXT: # ymm0 = mem[0,1,0,1] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_avx_vbroadcastf128_pd_256: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vbroadcastf128 (%rdi), %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x1a,0x07] +; X64-AVX512VL-NEXT: # ymm0 = mem[0,1,0,1] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.vbroadcastf128.pd.256(i8* %a0) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -114,16 +165,31 @@ declare <4 x double> @llvm.x86.avx.vbroadcastf128.pd.256(i8*) nounwind readonly define <8 x float> @test_x86_avx_vbroadcastf128_ps_256(i8* %a0) { -; X86-LABEL: test_x86_avx_vbroadcastf128_ps_256: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1] -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_avx_vbroadcastf128_ps_256: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vbroadcastf128 (%eax), %ymm0 # encoding: [0xc4,0xe2,0x7d,0x1a,0x00] +; X86-AVX-NEXT: # ymm0 = mem[0,1,0,1] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_avx_vbroadcastf128_ps_256: -; X64: # %bb.0: -; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1] -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_avx_vbroadcastf128_ps_256: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vbroadcastf128 (%eax), %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x1a,0x00] +; X86-AVX512VL-NEXT: # ymm0 = mem[0,1,0,1] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_avx_vbroadcastf128_ps_256: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vbroadcastf128 (%rdi), %ymm0 # encoding: [0xc4,0xe2,0x7d,0x1a,0x07] +; X64-AVX-NEXT: # ymm0 = mem[0,1,0,1] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_avx_vbroadcastf128_ps_256: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vbroadcastf128 (%rdi), %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0x1a,0x07] +; X64-AVX512VL-NEXT: # ymm0 = mem[0,1,0,1] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.vbroadcastf128.ps.256(i8* %a0) ; <<8 x float>> [#uses=1] ret <8 x float> %res } @@ -133,8 +199,9 @@ declare <8 x float> @llvm.x86.avx.vbroadcastf128.ps.256(i8*) nounwind readonly define <4 x double> @test_x86_avx_blend_pd_256(<4 x double> %a0, <4 x double> %a1) { ; CHECK-LABEL: test_x86_avx_blend_pd_256: ; CHECK: # %bb.0: -; CHECK-NEXT: vblendps {{.*#+}} ymm0 = ymm1[0,1,2,3,4,5],ymm0[6,7] -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vblendps $192, %ymm0, %ymm1, %ymm0 # encoding: [0xc4,0xe3,0x75,0x0c,0xc0,0xc0] +; CHECK-NEXT: # ymm0 = ymm1[0,1,2,3,4,5],ymm0[6,7] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.blend.pd.256(<4 x double> %a0, <4 x double> %a1, i32 7) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -144,8 +211,9 @@ declare <4 x double> @llvm.x86.avx.blend.pd.256(<4 x double>, <4 x double>, i32) define <8 x float> @test_x86_avx_blend_ps_256(<8 x float> %a0, <8 x float> %a1) { ; CHECK-LABEL: test_x86_avx_blend_ps_256: ; CHECK: # %bb.0: -; CHECK-NEXT: vblendps {{.*#+}} ymm0 = ymm1[0,1,2],ymm0[3,4,5,6,7] -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vblendps $7, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x0c,0xc1,0x07] +; CHECK-NEXT: # ymm0 = ymm1[0,1,2],ymm0[3,4,5,6,7] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.blend.ps.256(<8 x float> %a0, <8 x float> %a1, i32 7) ; <<8 x float>> [#uses=1] ret <8 x float> %res } @@ -155,8 +223,8 @@ declare <8 x float> @llvm.x86.avx.blend.ps.256(<8 x float>, <8 x float>, i32) no define <8 x float> @test_x86_avx_dp_ps_256(<8 x float> %a0, <8 x float> %a1) { ; CHECK-LABEL: test_x86_avx_dp_ps_256: ; CHECK: # %bb.0: -; CHECK-NEXT: vdpps $7, %ymm1, %ymm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vdpps $7, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x40,0xc1,0x07] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.dp.ps.256(<8 x float> %a0, <8 x float> %a1, i32 7) ; <<8 x float>> [#uses=1] ret <8 x float> %res } @@ -164,10 +232,17 @@ declare <8 x float> @llvm.x86.avx.dp.ps.256(<8 x float>, <8 x float>, i32) nounw define <2 x i64> @test_x86_sse2_psll_dq(<2 x i64> %a0) { -; CHECK-LABEL: test_x86_sse2_psll_dq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpslldq {{.*#+}} xmm0 = zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse2_psll_dq: +; AVX: # %bb.0: +; AVX-NEXT: vpslldq $1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0x73,0xf8,0x01] +; AVX-NEXT: # xmm0 = zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse2_psll_dq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpslldq $1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x73,0xf8,0x01] +; AVX512VL-NEXT: # xmm0 = zero,xmm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse2.psll.dq(<2 x i64> %a0, i32 8) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -175,10 +250,17 @@ declare <2 x i64> @llvm.x86.sse2.psll.dq(<2 x i64>, i32) nounwind readnone define <2 x i64> @test_x86_sse2_psrl_dq(<2 x i64> %a0) { -; CHECK-LABEL: test_x86_sse2_psrl_dq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpsrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse2_psrl_dq: +; AVX: # %bb.0: +; AVX-NEXT: vpsrldq $1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0x73,0xd8,0x01] +; AVX-NEXT: # xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse2_psrl_dq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpsrldq $1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x73,0xd8,0x01] +; AVX512VL-NEXT: # xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse2.psrl.dq(<2 x i64> %a0, i32 8) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -186,10 +268,17 @@ declare <2 x i64> @llvm.x86.sse2.psrl.dq(<2 x i64>, i32) nounwind readnone define <2 x double> @test_x86_sse41_blendpd(<2 x double> %a0, <2 x double> %a1) { -; CHECK-LABEL: test_x86_sse41_blendpd: -; CHECK: # %bb.0: -; CHECK-NEXT: vblendps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_blendpd: +; AVX: # %bb.0: +; AVX-NEXT: vblendps $3, %xmm0, %xmm1, %xmm0 # encoding: [0xc4,0xe3,0x71,0x0c,0xc0,0x03] +; AVX-NEXT: # xmm0 = xmm0[0,1],xmm1[2,3] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_blendpd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vmovsd %xmm0, %xmm1, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf3,0x10,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],xmm1[1] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.sse41.blendpd(<2 x double> %a0, <2 x double> %a1, i8 2) ; <<2 x double>> [#uses=1] ret <2 x double> %res } @@ -199,8 +288,9 @@ declare <2 x double> @llvm.x86.sse41.blendpd(<2 x double>, <2 x double>, i8) nou define <4 x float> @test_x86_sse41_blendps(<4 x float> %a0, <4 x float> %a1) { ; CHECK-LABEL: test_x86_sse41_blendps: ; CHECK: # %bb.0: -; CHECK-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0,1,2],xmm0[3] -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vblendps $8, %xmm0, %xmm1, %xmm0 # encoding: [0xc4,0xe3,0x71,0x0c,0xc0,0x08] +; CHECK-NEXT: # xmm0 = xmm1[0,1,2],xmm0[3] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x float> @llvm.x86.sse41.blendps(<4 x float> %a0, <4 x float> %a1, i8 7) ; <<4 x float>> [#uses=1] ret <4 x float> %res } @@ -210,8 +300,9 @@ declare <4 x float> @llvm.x86.sse41.blendps(<4 x float>, <4 x float>, i8) nounwi define <8 x i16> @test_x86_sse41_pblendw(<8 x i16> %a0, <8 x i16> %a1) { ; CHECK-LABEL: test_x86_sse41_pblendw: ; CHECK: # %bb.0: -; CHECK-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1,2],xmm0[3,4,5,6,7] -; CHECK-NEXT: ret{{[l|q]}} +; CHECK-NEXT: vpblendw $7, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0x79,0x0e,0xc1,0x07] +; CHECK-NEXT: # xmm0 = xmm1[0,1,2],xmm0[3,4,5,6,7] +; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i16> @llvm.x86.sse41.pblendw(<8 x i16> %a0, <8 x i16> %a1, i8 7) ; <<8 x i16>> [#uses=1] ret <8 x i16> %res } @@ -219,10 +310,15 @@ declare <8 x i16> @llvm.x86.sse41.pblendw(<8 x i16>, <8 x i16>, i8) nounwind rea define <4 x i32> @test_x86_sse41_pmovsxbd(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxbd: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxbd %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxbd: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxbd %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x21,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxbd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxbd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x21,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x i32> @llvm.x86.sse41.pmovsxbd(<16 x i8> %a0) ; <<4 x i32>> [#uses=1] ret <4 x i32> %res } @@ -230,10 +326,15 @@ declare <4 x i32> @llvm.x86.sse41.pmovsxbd(<16 x i8>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovsxbq(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxbq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxbq %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxbq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxbq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x22,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxbq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxbq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x22,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovsxbq(<16 x i8> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -241,10 +342,15 @@ declare <2 x i64> @llvm.x86.sse41.pmovsxbq(<16 x i8>) nounwind readnone define <8 x i16> @test_x86_sse41_pmovsxbw(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxbw: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxbw %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxbw: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxbw %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x20,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxbw: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxbw %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x20,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i16> @llvm.x86.sse41.pmovsxbw(<16 x i8> %a0) ; <<8 x i16>> [#uses=1] ret <8 x i16> %res } @@ -252,10 +358,15 @@ declare <8 x i16> @llvm.x86.sse41.pmovsxbw(<16 x i8>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovsxdq(<4 x i32> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxdq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxdq %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxdq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxdq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x25,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxdq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxdq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x25,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovsxdq(<4 x i32> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -263,10 +374,15 @@ declare <2 x i64> @llvm.x86.sse41.pmovsxdq(<4 x i32>) nounwind readnone define <4 x i32> @test_x86_sse41_pmovsxwd(<8 x i16> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxwd: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxwd %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxwd: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxwd %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x23,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxwd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxwd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x23,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16> %a0) ; <<4 x i32>> [#uses=1] ret <4 x i32> %res } @@ -274,10 +390,15 @@ declare <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovsxwq(<8 x i16> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovsxwq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovsxwq %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovsxwq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovsxwq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x24,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovsxwq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovsxwq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x24,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovsxwq(<8 x i16> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -285,10 +406,17 @@ declare <2 x i64> @llvm.x86.sse41.pmovsxwq(<8 x i16>) nounwind readnone define <4 x i32> @test_x86_sse41_pmovzxbd(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxbd: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxbd {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxbd: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxbd %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x31,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxbd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxbd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x31,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x i32> @llvm.x86.sse41.pmovzxbd(<16 x i8> %a0) ; <<4 x i32>> [#uses=1] ret <4 x i32> %res } @@ -296,10 +424,17 @@ declare <4 x i32> @llvm.x86.sse41.pmovzxbd(<16 x i8>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovzxbq(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxbq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxbq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxbq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxbq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x32,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxbq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxbq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x32,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovzxbq(<16 x i8> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -307,10 +442,17 @@ declare <2 x i64> @llvm.x86.sse41.pmovzxbq(<16 x i8>) nounwind readnone define <8 x i16> @test_x86_sse41_pmovzxbw(<16 x i8> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxbw: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxbw: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxbw %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x30,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxbw: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxbw %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x30,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i16> @llvm.x86.sse41.pmovzxbw(<16 x i8> %a0) ; <<8 x i16>> [#uses=1] ret <8 x i16> %res } @@ -318,10 +460,17 @@ declare <8 x i16> @llvm.x86.sse41.pmovzxbw(<16 x i8>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovzxdq(<4 x i32> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxdq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxdq {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxdq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxdq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x35,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxdq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxdq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x35,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovzxdq(<4 x i32> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -329,10 +478,17 @@ declare <2 x i64> @llvm.x86.sse41.pmovzxdq(<4 x i32>) nounwind readnone define <4 x i32> @test_x86_sse41_pmovzxwd(<8 x i16> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxwd: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxwd: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxwd %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x33,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxwd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxwd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x33,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x i32> @llvm.x86.sse41.pmovzxwd(<8 x i16> %a0) ; <<4 x i32>> [#uses=1] ret <4 x i32> %res } @@ -340,10 +496,17 @@ declare <4 x i32> @llvm.x86.sse41.pmovzxwd(<8 x i16>) nounwind readnone define <2 x i64> @test_x86_sse41_pmovzxwq(<8 x i16> %a0) { -; CHECK-LABEL: test_x86_sse41_pmovzxwq: -; CHECK: # %bb.0: -; CHECK-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse41_pmovzxwq: +; AVX: # %bb.0: +; AVX-NEXT: vpmovzxwq %xmm0, %xmm0 # encoding: [0xc4,0xe2,0x79,0x34,0xc0] +; AVX-NEXT: # xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse41_pmovzxwq: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpmovzxwq %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0x34,0xc0] +; AVX512VL-NEXT: # xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x i64> @llvm.x86.sse41.pmovzxwq(<8 x i16> %a0) ; <<2 x i64>> [#uses=1] ret <2 x i64> %res } @@ -351,10 +514,15 @@ declare <2 x i64> @llvm.x86.sse41.pmovzxwq(<8 x i16>) nounwind readnone define <2 x double> @test_x86_sse2_cvtdq2pd(<4 x i32> %a0) { -; CHECK-LABEL: test_x86_sse2_cvtdq2pd: -; CHECK: # %bb.0: -; CHECK-NEXT: vcvtdq2pd %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse2_cvtdq2pd: +; AVX: # %bb.0: +; AVX-NEXT: vcvtdq2pd %xmm0, %xmm0 # encoding: [0xc5,0xfa,0xe6,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse2_cvtdq2pd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vcvtdq2pd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xfa,0xe6,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.sse2.cvtdq2pd(<4 x i32> %a0) ; <<2 x double>> [#uses=1] ret <2 x double> %res } @@ -362,10 +530,15 @@ declare <2 x double> @llvm.x86.sse2.cvtdq2pd(<4 x i32>) nounwind readnone define <4 x double> @test_x86_avx_cvtdq2_pd_256(<4 x i32> %a0) { -; CHECK-LABEL: test_x86_avx_cvtdq2_pd_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vcvtdq2pd %xmm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_cvtdq2_pd_256: +; AVX: # %bb.0: +; AVX-NEXT: vcvtdq2pd %xmm0, %ymm0 # encoding: [0xc5,0xfe,0xe6,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_cvtdq2_pd_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vcvtdq2pd %xmm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfe,0xe6,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.cvtdq2.pd.256(<4 x i32> %a0) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -373,10 +546,15 @@ declare <4 x double> @llvm.x86.avx.cvtdq2.pd.256(<4 x i32>) nounwind readnone define <2 x double> @test_x86_sse2_cvtps2pd(<4 x float> %a0) { -; CHECK-LABEL: test_x86_sse2_cvtps2pd: -; CHECK: # %bb.0: -; CHECK-NEXT: vcvtps2pd %xmm0, %xmm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_sse2_cvtps2pd: +; AVX: # %bb.0: +; AVX-NEXT: vcvtps2pd %xmm0, %xmm0 # encoding: [0xc5,0xf8,0x5a,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_sse2_cvtps2pd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vcvtps2pd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.sse2.cvtps2pd(<4 x float> %a0) ; <<2 x double>> [#uses=1] ret <2 x double> %res } @@ -384,10 +562,15 @@ declare <2 x double> @llvm.x86.sse2.cvtps2pd(<4 x float>) nounwind readnone define <4 x double> @test_x86_avx_cvt_ps2_pd_256(<4 x float> %a0) { -; CHECK-LABEL: test_x86_avx_cvt_ps2_pd_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vcvtps2pd %xmm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_cvt_ps2_pd_256: +; AVX: # %bb.0: +; AVX-NEXT: vcvtps2pd %xmm0, %ymm0 # encoding: [0xc5,0xfc,0x5a,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_cvt_ps2_pd_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vcvtps2pd %xmm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x5a,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.cvt.ps2.pd.256(<4 x float> %a0) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -396,20 +579,35 @@ declare <4 x double> @llvm.x86.avx.cvt.ps2.pd.256(<4 x float>) nounwind readnone define void @test_x86_sse2_storeu_dq(i8* %a0, <16 x i8> %a1) { ; add operation forces the execution domain. -; X86-LABEL: test_x86_sse2_storeu_dq: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 -; X86-NEXT: vpsubb %xmm1, %xmm0, %xmm0 -; X86-NEXT: vmovdqu %xmm0, (%eax) -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_sse2_storeu_dq: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x76,0xc9] +; X86-AVX-NEXT: vpsubb %xmm1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0xf8,0xc1] +; X86-AVX-NEXT: vmovdqu %xmm0, (%eax) # encoding: [0xc5,0xfa,0x7f,0x00] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_sse2_storeu_dq: -; X64: # %bb.0: -; X64-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 -; X64-NEXT: vpsubb %xmm1, %xmm0, %xmm0 -; X64-NEXT: vmovdqu %xmm0, (%rdi) -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_sse2_storeu_dq: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x76,0xc9] +; X86-AVX512VL-NEXT: vpsubb %xmm1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xf8,0xc1] +; X86-AVX512VL-NEXT: vmovdqu %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xfa,0x7f,0x00] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_sse2_storeu_dq: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x76,0xc9] +; X64-AVX-NEXT: vpsubb %xmm1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0xf8,0xc1] +; X64-AVX-NEXT: vmovdqu %xmm0, (%rdi) # encoding: [0xc5,0xfa,0x7f,0x07] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_sse2_storeu_dq: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x76,0xc9] +; X64-AVX512VL-NEXT: vpsubb %xmm1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0xf8,0xc1] +; X64-AVX512VL-NEXT: vmovdqu %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xfa,0x7f,0x07] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %a2 = add <16 x i8> %a1, call void @llvm.x86.sse2.storeu.dq(i8* %a0, <16 x i8> %a2) ret void @@ -419,22 +617,49 @@ declare void @llvm.x86.sse2.storeu.dq(i8*, <16 x i8>) nounwind define void @test_x86_sse2_storeu_pd(i8* %a0, <2 x double> %a1) { ; fadd operation forces the execution domain. -; X86-LABEL: test_x86_sse2_storeu_pd: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; X86-NEXT: vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0] -; X86-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; X86-NEXT: vmovupd %xmm0, (%eax) -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_sse2_storeu_pd: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x57,0xc9] +; X86-AVX-NEXT: vmovhpd {{\.LCPI.*}}, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x16,0x0d,A,A,A,A] +; X86-AVX-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}, kind: FK_Data_4 +; X86-AVX-NEXT: # xmm1 = xmm1[0],mem[0] +; X86-AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0x58,0xc1] +; X86-AVX-NEXT: vmovupd %xmm0, (%eax) # encoding: [0xc5,0xf9,0x11,0x00] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_sse2_storeu_pd: -; X64: # %bb.0: -; X64-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; X64-NEXT: vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0] -; X64-NEXT: vaddpd %xmm1, %xmm0, %xmm0 -; X64-NEXT: vmovupd %xmm0, (%rdi) -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_sse2_storeu_pd: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vmovsd {{\.LCPI.*}}, %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xfb,0x10,0x0d,A,A,A,A] +; X86-AVX512VL-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}, kind: FK_Data_4 +; X86-AVX512VL-NEXT: # xmm1 = mem[0],zero +; X86-AVX512VL-NEXT: vpslldq $8, %xmm1, %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x73,0xf9,0x08] +; X86-AVX512VL-NEXT: # xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7] +; X86-AVX512VL-NEXT: vaddpd %xmm1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x58,0xc1] +; X86-AVX512VL-NEXT: vmovupd %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x11,0x00] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_sse2_storeu_pd: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x57,0xc9] +; X64-AVX-NEXT: vmovhpd {{.*}}(%rip), %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x16,0x0d,A,A,A,A] +; X64-AVX-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}-4, kind: reloc_riprel_4byte +; X64-AVX-NEXT: # xmm1 = xmm1[0],mem[0] +; X64-AVX-NEXT: vaddpd %xmm1, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0x58,0xc1] +; X64-AVX-NEXT: vmovupd %xmm0, (%rdi) # encoding: [0xc5,0xf9,0x11,0x07] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_sse2_storeu_pd: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vmovsd {{.*}}(%rip), %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xfb,0x10,0x0d,A,A,A,A] +; X64-AVX512VL-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}-4, kind: reloc_riprel_4byte +; X64-AVX512VL-NEXT: # xmm1 = mem[0],zero +; X64-AVX512VL-NEXT: vpslldq $8, %xmm1, %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x73,0xf9,0x08] +; X64-AVX512VL-NEXT: # xmm1 = zero,zero,zero,zero,zero,zero,zero,zero,xmm1[0,1,2,3,4,5,6,7] +; X64-AVX512VL-NEXT: vaddpd %xmm1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x58,0xc1] +; X64-AVX512VL-NEXT: vmovupd %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xf9,0x11,0x07] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %a2 = fadd <2 x double> %a1, call void @llvm.x86.sse2.storeu.pd(i8* %a0, <2 x double> %a2) ret void @@ -443,16 +668,27 @@ declare void @llvm.x86.sse2.storeu.pd(i8*, <2 x double>) nounwind define void @test_x86_sse_storeu_ps(i8* %a0, <4 x float> %a1) { -; X86-LABEL: test_x86_sse_storeu_ps: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vmovups %xmm0, (%eax) -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_sse_storeu_ps: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vmovups %xmm0, (%eax) # encoding: [0xc5,0xf8,0x11,0x00] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_sse_storeu_ps: -; X64: # %bb.0: -; X64-NEXT: vmovups %xmm0, (%rdi) -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_sse_storeu_ps: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vmovups %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x11,0x00] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_sse_storeu_ps: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vmovups %xmm0, (%rdi) # encoding: [0xc5,0xf8,0x11,0x07] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_sse_storeu_ps: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vmovups %xmm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x11,0x07] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] call void @llvm.x86.sse.storeu.ps(i8* %a0, <4 x float> %a1) ret void } @@ -462,28 +698,45 @@ declare void @llvm.x86.sse.storeu.ps(i8*, <4 x float>) nounwind define void @test_x86_avx_storeu_dq_256(i8* %a0, <32 x i8> %a1) { ; FIXME: unfortunately the execution domain fix pass changes this to vmovups and its hard to force with no 256-bit integer instructions ; add operation forces the execution domain. -; X86-LABEL: test_x86_avx_storeu_dq_256: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vextractf128 $1, %ymm0, %xmm1 -; X86-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 -; X86-NEXT: vpsubb %xmm2, %xmm1, %xmm1 -; X86-NEXT: vpsubb %xmm2, %xmm0, %xmm0 -; X86-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; X86-NEXT: vmovups %ymm0, (%eax) -; X86-NEXT: vzeroupper -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_avx_storeu_dq_256: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 # encoding: [0xc4,0xe3,0x7d,0x19,0xc1,0x01] +; X86-AVX-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0x76,0xd2] +; X86-AVX-NEXT: vpsubb %xmm2, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0xf8,0xca] +; X86-AVX-NEXT: vpsubb %xmm2, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0xf8,0xc2] +; X86-AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; X86-AVX-NEXT: vmovups %ymm0, (%eax) # encoding: [0xc5,0xfc,0x11,0x00] +; X86-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_avx_storeu_dq_256: -; X64: # %bb.0: -; X64-NEXT: vextractf128 $1, %ymm0, %xmm1 -; X64-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 -; X64-NEXT: vpsubb %xmm2, %xmm1, %xmm1 -; X64-NEXT: vpsubb %xmm2, %xmm0, %xmm0 -; X64-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 -; X64-NEXT: vmovups %ymm0, (%rdi) -; X64-NEXT: vzeroupper -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_avx_storeu_dq_256: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1 # encoding: [0xc5,0xf5,0x76,0xc9] +; X86-AVX512VL-NEXT: vpsubb %ymm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xf8,0xc1] +; X86-AVX512VL-NEXT: vmovdqu %ymm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xfe,0x7f,0x00] +; X86-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_avx_storeu_dq_256: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vextractf128 $1, %ymm0, %xmm1 # encoding: [0xc4,0xe3,0x7d,0x19,0xc1,0x01] +; X64-AVX-NEXT: vpcmpeqd %xmm2, %xmm2, %xmm2 # encoding: [0xc5,0xe9,0x76,0xd2] +; X64-AVX-NEXT: vpsubb %xmm2, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0xf8,0xca] +; X64-AVX-NEXT: vpsubb %xmm2, %xmm0, %xmm0 # encoding: [0xc5,0xf9,0xf8,0xc2] +; X64-AVX-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x18,0xc1,0x01] +; X64-AVX-NEXT: vmovups %ymm0, (%rdi) # encoding: [0xc5,0xfc,0x11,0x07] +; X64-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_avx_storeu_dq_256: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1 # encoding: [0xc5,0xf5,0x76,0xc9] +; X64-AVX512VL-NEXT: vpsubb %ymm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0xf8,0xc1] +; X64-AVX512VL-NEXT: vmovdqu %ymm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xfe,0x7f,0x07] +; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %a2 = add <32 x i8> %a1, call void @llvm.x86.avx.storeu.dq.256(i8* %a0, <32 x i8> %a2) ret void @@ -493,22 +746,39 @@ declare void @llvm.x86.avx.storeu.dq.256(i8*, <32 x i8>) nounwind define void @test_x86_avx_storeu_pd_256(i8* %a0, <4 x double> %a1) { ; add operation forces the execution domain. -; X86-LABEL: test_x86_avx_storeu_pd_256: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; X86-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; X86-NEXT: vmovupd %ymm0, (%eax) -; X86-NEXT: vzeroupper -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_avx_storeu_pd_256: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x57,0xc9] +; X86-AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # encoding: [0xc5,0xfd,0x58,0xc1] +; X86-AVX-NEXT: vmovupd %ymm0, (%eax) # encoding: [0xc5,0xfd,0x11,0x00] +; X86-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_avx_storeu_pd_256: -; X64: # %bb.0: -; X64-NEXT: vxorpd %xmm1, %xmm1, %xmm1 -; X64-NEXT: vaddpd %ymm1, %ymm0, %ymm0 -; X64-NEXT: vmovupd %ymm0, (%rdi) -; X64-NEXT: vzeroupper -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_avx_storeu_pd_256: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x57,0xc9] +; X86-AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x58,0xc1] +; X86-AVX512VL-NEXT: vmovupd %ymm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x11,0x00] +; X86-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_avx_storeu_pd_256: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # encoding: [0xc5,0xf1,0x57,0xc9] +; X64-AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # encoding: [0xc5,0xfd,0x58,0xc1] +; X64-AVX-NEXT: vmovupd %ymm0, (%rdi) # encoding: [0xc5,0xfd,0x11,0x07] +; X64-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_avx_storeu_pd_256: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vxorpd %xmm1, %xmm1, %xmm1 # EVEX TO VEX Compression encoding: [0xc5,0xf1,0x57,0xc9] +; X64-AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x58,0xc1] +; X64-AVX512VL-NEXT: vmovupd %ymm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xfd,0x11,0x07] +; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] %a2 = fadd <4 x double> %a1, call void @llvm.x86.avx.storeu.pd.256(i8* %a0, <4 x double> %a2) ret void @@ -517,18 +787,31 @@ declare void @llvm.x86.avx.storeu.pd.256(i8*, <4 x double>) nounwind define void @test_x86_avx_storeu_ps_256(i8* %a0, <8 x float> %a1) { -; X86-LABEL: test_x86_avx_storeu_ps_256: -; X86: # %bb.0: -; X86-NEXT: movl {{[0-9]+}}(%esp), %eax -; X86-NEXT: vmovups %ymm0, (%eax) -; X86-NEXT: vzeroupper -; X86-NEXT: retl +; X86-AVX-LABEL: test_x86_avx_storeu_ps_256: +; X86-AVX: # %bb.0: +; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX-NEXT: vmovups %ymm0, (%eax) # encoding: [0xc5,0xfc,0x11,0x00] +; X86-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX-NEXT: retl # encoding: [0xc3] ; -; X64-LABEL: test_x86_avx_storeu_ps_256: -; X64: # %bb.0: -; X64-NEXT: vmovups %ymm0, (%rdi) -; X64-NEXT: vzeroupper -; X64-NEXT: retq +; X86-AVX512VL-LABEL: test_x86_avx_storeu_ps_256: +; X86-AVX512VL: # %bb.0: +; X86-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04] +; X86-AVX512VL-NEXT: vmovups %ymm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x11,0x00] +; X86-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X86-AVX512VL-NEXT: retl # encoding: [0xc3] +; +; X64-AVX-LABEL: test_x86_avx_storeu_ps_256: +; X64-AVX: # %bb.0: +; X64-AVX-NEXT: vmovups %ymm0, (%rdi) # encoding: [0xc5,0xfc,0x11,0x07] +; X64-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX-NEXT: retq # encoding: [0xc3] +; +; X64-AVX512VL-LABEL: test_x86_avx_storeu_ps_256: +; X64-AVX512VL: # %bb.0: +; X64-AVX512VL-NEXT: vmovups %ymm0, (%rdi) # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x11,0x07] +; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; X64-AVX512VL-NEXT: retq # encoding: [0xc3] call void @llvm.x86.avx.storeu.ps.256(i8* %a0, <8 x float> %a1) ret void } @@ -536,10 +819,17 @@ declare void @llvm.x86.avx.storeu.ps.256(i8*, <8 x float>) nounwind define <2 x double> @test_x86_avx_vpermil_pd(<2 x double> %a0) { -; CHECK-LABEL: test_x86_avx_vpermil_pd: -; CHECK: # %bb.0: -; CHECK-NEXT: vpermilpd {{.*#+}} xmm0 = xmm0[1,0] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vpermil_pd: +; AVX: # %bb.0: +; AVX-NEXT: vpermilpd $1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0x79,0x05,0xc0,0x01] +; AVX-NEXT: # xmm0 = xmm0[1,0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vpermil_pd: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpermilpd $1, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x05,0xc0,0x01] +; AVX512VL-NEXT: # xmm0 = xmm0[1,0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <2 x double> @llvm.x86.avx.vpermil.pd(<2 x double> %a0, i8 1) ; <<2 x double>> [#uses=1] ret <2 x double> %res } @@ -547,10 +837,17 @@ declare <2 x double> @llvm.x86.avx.vpermil.pd(<2 x double>, i8) nounwind readnon define <4 x double> @test_x86_avx_vpermil_pd_256(<4 x double> %a0) { -; CHECK-LABEL: test_x86_avx_vpermil_pd_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,2] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vpermil_pd_256: +; AVX: # %bb.0: +; AVX-NEXT: vpermilpd $7, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x05,0xc0,0x07] +; AVX-NEXT: # ymm0 = ymm0[1,1,3,2] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vpermil_pd_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpermilpd $7, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x05,0xc0,0x07] +; AVX512VL-NEXT: # ymm0 = ymm0[1,1,3,2] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.vpermil.pd.256(<4 x double> %a0, i8 7) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -558,10 +855,17 @@ declare <4 x double> @llvm.x86.avx.vpermil.pd.256(<4 x double>, i8) nounwind rea define <4 x float> @test_x86_avx_vpermil_ps(<4 x float> %a0) { -; CHECK-LABEL: test_x86_avx_vpermil_ps: -; CHECK: # %bb.0: -; CHECK-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,1,0,0] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vpermil_ps: +; AVX: # %bb.0: +; AVX-NEXT: vpermilps $7, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0x79,0x04,0xc0,0x07] +; AVX-NEXT: # xmm0 = xmm0[3,1,0,0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vpermil_ps: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpermilps $7, %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x79,0x04,0xc0,0x07] +; AVX512VL-NEXT: # xmm0 = xmm0[3,1,0,0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x float> @llvm.x86.avx.vpermil.ps(<4 x float> %a0, i8 7) ; <<4 x float>> [#uses=1] ret <4 x float> %res } @@ -569,10 +873,17 @@ declare <4 x float> @llvm.x86.avx.vpermil.ps(<4 x float>, i8) nounwind readnone define <8 x float> @test_x86_avx_vpermil_ps_256(<8 x float> %a0) { -; CHECK-LABEL: test_x86_avx_vpermil_ps_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[3,1,0,0,7,5,4,4] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vpermil_ps_256: +; AVX: # %bb.0: +; AVX-NEXT: vpermilps $7, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0x7d,0x04,0xc0,0x07] +; AVX-NEXT: # ymm0 = ymm0[3,1,0,0,7,5,4,4] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vpermil_ps_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vpermilps $7, %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x7d,0x04,0xc0,0x07] +; AVX512VL-NEXT: # ymm0 = ymm0[3,1,0,0,7,5,4,4] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.vpermil.ps.256(<8 x float> %a0, i8 7) ; <<8 x float>> [#uses=1] ret <8 x float> %res } @@ -580,10 +891,17 @@ declare <8 x float> @llvm.x86.avx.vpermil.ps.256(<8 x float>, i8) nounwind readn define <4 x double> @test_x86_avx_vperm2f128_pd_256(<4 x double> %a0, <4 x double> %a1) { -; CHECK-LABEL: test_x86_avx_vperm2f128_pd_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vperm2f128_pd_256: +; AVX: # %bb.0: +; AVX-NEXT: vperm2f128 $33, %ymm0, %ymm1, %ymm0 # encoding: [0xc4,0xe3,0x75,0x06,0xc0,0x21] +; AVX-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vperm2f128_pd_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vperm2f128 $33, %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x75,0x06,0xc0,0x21] +; AVX512VL-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <4 x double> @llvm.x86.avx.vperm2f128.pd.256(<4 x double> %a0, <4 x double> %a1, i8 3) ; <<4 x double>> [#uses=1] ret <4 x double> %res } @@ -591,10 +909,17 @@ declare <4 x double> @llvm.x86.avx.vperm2f128.pd.256(<4 x double>, <4 x double>, define <8 x float> @test_x86_avx_vperm2f128_ps_256(<8 x float> %a0, <8 x float> %a1) { -; CHECK-LABEL: test_x86_avx_vperm2f128_ps_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vperm2f128_ps_256: +; AVX: # %bb.0: +; AVX-NEXT: vperm2f128 $33, %ymm0, %ymm1, %ymm0 # encoding: [0xc4,0xe3,0x75,0x06,0xc0,0x21] +; AVX-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vperm2f128_ps_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vperm2f128 $33, %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x75,0x06,0xc0,0x21] +; AVX512VL-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.vperm2f128.ps.256(<8 x float> %a0, <8 x float> %a1, i8 3) ; <<8 x float>> [#uses=1] ret <8 x float> %res } @@ -602,10 +927,17 @@ declare <8 x float> @llvm.x86.avx.vperm2f128.ps.256(<8 x float>, <8 x float>, i8 define <8 x i32> @test_x86_avx_vperm2f128_si_256(<8 x i32> %a0, <8 x i32> %a1) { -; CHECK-LABEL: test_x86_avx_vperm2f128_si_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm1[2,3],ymm0[0,1] -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_vperm2f128_si_256: +; AVX: # %bb.0: +; AVX-NEXT: vperm2f128 $33, %ymm0, %ymm1, %ymm0 # encoding: [0xc4,0xe3,0x75,0x06,0xc0,0x21] +; AVX-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_vperm2f128_si_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vperm2i128 $33, %ymm0, %ymm1, %ymm0 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0x75,0x46,0xc0,0x21] +; AVX512VL-NEXT: # ymm0 = ymm1[2,3],ymm0[0,1] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x i32> @llvm.x86.avx.vperm2f128.si.256(<8 x i32> %a0, <8 x i32> %a1, i8 3) ; <<8 x i32>> [#uses=1] ret <8 x i32> %res } @@ -613,10 +945,15 @@ declare <8 x i32> @llvm.x86.avx.vperm2f128.si.256(<8 x i32>, <8 x i32>, i8) noun define <8 x float> @test_x86_avx_cvtdq2_ps_256(<8 x i32> %a0) { -; CHECK-LABEL: test_x86_avx_cvtdq2_ps_256: -; CHECK: # %bb.0: -; CHECK-NEXT: vcvtdq2ps %ymm0, %ymm0 -; CHECK-NEXT: ret{{[l|q]}} +; AVX-LABEL: test_x86_avx_cvtdq2_ps_256: +; AVX: # %bb.0: +; AVX-NEXT: vcvtdq2ps %ymm0, %ymm0 # encoding: [0xc5,0xfc,0x5b,0xc0] +; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3] +; +; AVX512VL-LABEL: test_x86_avx_cvtdq2_ps_256: +; AVX512VL: # %bb.0: +; AVX512VL-NEXT: vcvtdq2ps %ymm0, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x5b,0xc0] +; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3] %res = call <8 x float> @llvm.x86.avx.cvtdq2.ps.256(<8 x i32> %a0) ; <<8 x float>> [#uses=1] ret <8 x float> %res } diff --git a/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll b/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll index 93d5da65c3b2..ace9ae24f278 100644 --- a/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll +++ b/llvm/test/CodeGen/X86/avx-intrinsics-x86.ll @@ -1,8 +1,8 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=avx,pclmul -show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=X86 --check-prefix=X86-AVX -; RUN: llc < %s -mtriple=i686-unknown-unknown -mcpu=skx -show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512VL --check-prefix=X86 --check-prefix=X86-AVX512VL -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx,pclmul -show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=X64 --check-prefix=X64-AVX -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mcpu=skx -show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512VL --check-prefix=X64 --check-prefix=X64-AVX512VL +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+pclmul,+avx -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86,AVX,X86-AVX +; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+pclmul,+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86,AVX512VL,X86-AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+pclmul,+avx -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64,AVX,X64-AVX +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+pclmul,+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64,AVX512VL,X64-AVX512VL define <4 x double> @test_x86_avx_addsub_pd_256(<4 x double> %a0, <4 x double> %a1) { ; CHECK-LABEL: test_x86_avx_addsub_pd_256: diff --git a/llvm/test/CodeGen/X86/avx-intrinsics-x86_64.ll b/llvm/test/CodeGen/X86/avx-intrinsics-x86_64.ll index c261a536b6e5..3a68e71fe3ef 100644 --- a/llvm/test/CodeGen/X86/avx-intrinsics-x86_64.ll +++ b/llvm/test/CodeGen/X86/avx-intrinsics-x86_64.ll @@ -1,22 +1,24 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=CHECK --check-prefix=AVX -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512VL +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,AVX +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=CHECK,AVX512VL define <4 x double> @test_x86_avx_vzeroall(<4 x double> %a, <4 x double> %b) { ; AVX-LABEL: test_x86_avx_vzeroall: ; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # encoding: [0xc5,0xfd,0x58,0xc1] ; AVX-NEXT: vmovupd %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill -; AVX-NEXT: vzeroall +; AVX-NEXT: # encoding: [0xc5,0xfd,0x11,0x44,0x24,0xc8] +; AVX-NEXT: vzeroall # encoding: [0xc5,0xfc,0x77] ; AVX-NEXT: vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload -; AVX-NEXT: retq +; AVX-NEXT: # encoding: [0xc5,0xfc,0x10,0x44,0x24,0xc8] +; AVX-NEXT: retq # encoding: [0xc3] ; ; AVX512VL-LABEL: test_x86_avx_vzeroall: ; AVX512VL: # %bb.0: -; AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm16 -; AVX512VL-NEXT: vzeroall -; AVX512VL-NEXT: vmovapd %ymm16, %ymm0 -; AVX512VL-NEXT: retq +; AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm16 # encoding: [0x62,0xe1,0xfd,0x28,0x58,0xc1] +; AVX512VL-NEXT: vzeroall # encoding: [0xc5,0xfc,0x77] +; AVX512VL-NEXT: vmovapd %ymm16, %ymm0 # encoding: [0x62,0xb1,0xfd,0x28,0x28,0xc0] +; AVX512VL-NEXT: retq # encoding: [0xc3] %c = fadd <4 x double> %a, %b call void @llvm.x86.avx.vzeroall() ret <4 x double> %c @@ -26,18 +28,20 @@ declare void @llvm.x86.avx.vzeroall() nounwind define <4 x double> @test_x86_avx_vzeroupper(<4 x double> %a, <4 x double> %b) { ; AVX-LABEL: test_x86_avx_vzeroupper: ; AVX: # %bb.0: -; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 +; AVX-NEXT: vaddpd %ymm1, %ymm0, %ymm0 # encoding: [0xc5,0xfd,0x58,0xc1] ; AVX-NEXT: vmovupd %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill -; AVX-NEXT: vzeroupper +; AVX-NEXT: # encoding: [0xc5,0xfd,0x11,0x44,0x24,0xc8] +; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] ; AVX-NEXT: vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload -; AVX-NEXT: retq +; AVX-NEXT: # encoding: [0xc5,0xfc,0x10,0x44,0x24,0xc8] +; AVX-NEXT: retq # encoding: [0xc3] ; ; AVX512VL-LABEL: test_x86_avx_vzeroupper: ; AVX512VL: # %bb.0: -; AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm16 -; AVX512VL-NEXT: vzeroupper -; AVX512VL-NEXT: vmovapd %ymm16, %ymm0 -; AVX512VL-NEXT: retq +; AVX512VL-NEXT: vaddpd %ymm1, %ymm0, %ymm16 # encoding: [0x62,0xe1,0xfd,0x28,0x58,0xc1] +; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77] +; AVX512VL-NEXT: vmovapd %ymm16, %ymm0 # encoding: [0x62,0xb1,0xfd,0x28,0x28,0xc0] +; AVX512VL-NEXT: retq # encoding: [0xc3] %c = fadd <4 x double> %a, %b call void @llvm.x86.avx.vzeroupper() ret <4 x double> %c