-; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=+avx2 < %s | FileCheck %s
+; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx2 < %s | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-unknown"
;CHECK-LABEL: stack_fold_broadcastsd_ymm
;CHECK: vbroadcastsd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <4 x double> @llvm.x86.avx2.vbroadcast.sd.pd.256(<2 x double> %a0)
+ %2 = shufflevector <2 x double> %a0, <2 x double> undef, <4 x i32> zeroinitializer
; fadd forces execution domain
%3 = fadd <4 x double> %2, <double 0x0, double 0x0, double 0x0, double 0x0>
ret <4 x double> %3
;CHECK-LABEL: stack_fold_broadcastss
;CHECK: vbroadcastss {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <4 x float> @llvm.x86.avx2.vbroadcast.ss.ps(<4 x float> %a0)
+ %2 = shufflevector <4 x float> %a0, <4 x float> undef, <4 x i32> zeroinitializer
; fadd forces execution domain
%3 = fadd <4 x float> %2, <float 0x0, float 0x0, float 0x0, float 0x0>
ret <4 x float> %3
;CHECK-LABEL: stack_fold_broadcastss_ymm
;CHECK: vbroadcastss {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <8 x float> @llvm.x86.avx2.vbroadcast.ss.ps.256(<4 x float> %a0)
+ %2 = shufflevector <4 x float> %a0, <4 x float> undef, <8 x i32> zeroinitializer
; fadd forces execution domain
%3 = fadd <8 x float> %2, <float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0, float 0x0>
ret <8 x float> %3
;CHECK-LABEL: stack_fold_pbroadcastb
;CHECK: vpbroadcastb {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <16 x i8> @llvm.x86.avx2.pbroadcastb.128(<16 x i8> %a0)
+ %2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <16 x i32> zeroinitializer
ret <16 x i8> %2
}
-declare <16 x i8> @llvm.x86.avx2.pbroadcastb.128(<16 x i8>) nounwind readonly
define <32 x i8> @stack_fold_pbroadcastb_ymm(<16 x i8> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastb_ymm
;CHECK: vpbroadcastb {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <32 x i8> @llvm.x86.avx2.pbroadcastb.256(<16 x i8> %a0)
+ %2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <32 x i32> zeroinitializer
ret <32 x i8> %2
}
-declare <32 x i8> @llvm.x86.avx2.pbroadcastb.256(<16 x i8>) nounwind readonly
define <4 x i32> @stack_fold_pbroadcastd(<4 x i32> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastd
;CHECK: vpbroadcastd {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <4 x i32> @llvm.x86.avx2.pbroadcastd.128(<4 x i32> %a0)
+ %2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <4 x i32> zeroinitializer
; add forces execution domain
%3 = add <4 x i32> %2, <i32 1, i32 1, i32 1, i32 1>
ret <4 x i32> %3
}
-declare <4 x i32> @llvm.x86.avx2.pbroadcastd.128(<4 x i32>) nounwind readonly
define <8 x i32> @stack_fold_pbroadcastd_ymm(<4 x i32> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastd_ymm
;CHECK: vpbroadcastd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <8 x i32> @llvm.x86.avx2.pbroadcastd.256(<4 x i32> %a0)
+ %2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <8 x i32> zeroinitializer
; add forces execution domain
%3 = add <8 x i32> %2, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
ret <8 x i32> %3
}
-declare <8 x i32> @llvm.x86.avx2.pbroadcastd.256(<4 x i32>) nounwind readonly
define <2 x i64> @stack_fold_pbroadcastq(<2 x i64> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastq
;CHECK: vpbroadcastq {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <2 x i64> @llvm.x86.avx2.pbroadcastq.128(<2 x i64> %a0)
+ %2 = shufflevector <2 x i64> %a0, <2 x i64> undef, <2 x i32> zeroinitializer
; add forces execution domain
%3 = add <2 x i64> %2, <i64 1, i64 1>
ret <2 x i64> %3
}
-declare <2 x i64> @llvm.x86.avx2.pbroadcastq.128(<2 x i64>) nounwind readonly
define <4 x i64> @stack_fold_pbroadcastq_ymm(<2 x i64> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastq_ymm
;CHECK: vpbroadcastq {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <4 x i64> @llvm.x86.avx2.pbroadcastq.256(<2 x i64> %a0)
+ %2 = shufflevector <2 x i64> %a0, <2 x i64> undef, <4 x i32> zeroinitializer
; add forces execution domain
%3 = add <4 x i64> %2, <i64 1, i64 1, i64 1, i64 1>
ret <4 x i64> %3
}
-declare <4 x i64> @llvm.x86.avx2.pbroadcastq.256(<2 x i64>) nounwind readonly
define <8 x i16> @stack_fold_pbroadcastw(<8 x i16> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastw
;CHECK: vpbroadcastw {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <8 x i16> @llvm.x86.avx2.pbroadcastw.128(<8 x i16> %a0)
+ %2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> zeroinitializer
ret <8 x i16> %2
}
-declare <8 x i16> @llvm.x86.avx2.pbroadcastw.128(<8 x i16>) nounwind readonly
define <16 x i16> @stack_fold_pbroadcastw_ymm(<8 x i16> %a0) {
;CHECK-LABEL: stack_fold_pbroadcastw_ymm
;CHECK: vpbroadcastw {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <16 x i16> @llvm.x86.avx2.pbroadcastw.256(<8 x i16> %a0)
+ %2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <16 x i32> zeroinitializer
ret <16 x i16> %2
}
-declare <16 x i16> @llvm.x86.avx2.pbroadcastw.256(<8 x i16>) nounwind readonly
define <32 x i8> @stack_fold_pcmpeqb(<32 x i8> %a0, <32 x i8> %a1) {
;CHECK-LABEL: stack_fold_pcmpeqb
define <4 x double> @stack_fold_permpd(<4 x double> %a0) {
;CHECK-LABEL: stack_fold_permpd
- ;CHECK: vpermpd $255, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+ ;CHECK: vpermpd $235, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = shufflevector <4 x double> %a0, <4 x double> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
+ %2 = shufflevector <4 x double> %a0, <4 x double> undef, <4 x i32> <i32 3, i32 2, i32 2, i32 3>
; fadd forces execution domain
%3 = fadd <4 x double> %2, <double 0x0, double 0x0, double 0x0, double 0x0>
ret <4 x double> %3
}
-define <8 x float> @stack_fold_permps(<8 x float> %a0, <8 x float> %a1) {
+define <8 x float> @stack_fold_permps(<8 x i32> %a0, <8 x float> %a1) {
;CHECK-LABEL: stack_fold_permps
;CHECK: vpermps {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a1, <8 x float> %a0)
+ %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a1, <8 x i32> %a0)
ret <8 x float> %2
}
-declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x float>) nounwind readonly
+declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x i32>) nounwind readonly
define <4 x i64> @stack_fold_permq(<4 x i64> %a0) {
;CHECK-LABEL: stack_fold_permq
- ;CHECK: vpermq $255, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+ ;CHECK: vpermq $235, {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
+ %2 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 2, i32 3>
; add forces execution domain
%3 = add <4 x i64> %2, <i64 1, i64 1, i64 1, i64 1>
ret <4 x i64> %3
;CHECK-LABEL: stack_fold_pmovsxbw
;CHECK: vpmovsxbw {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <16 x i16> @llvm.x86.avx2.pmovsxbw(<16 x i8> %a0)
+ %2 = sext <16 x i8> %a0 to <16 x i16>
ret <16 x i16> %2
}
-declare <16 x i16> @llvm.x86.avx2.pmovsxbw(<16 x i8>) nounwind readnone
define <4 x i64> @stack_fold_pmovsxdq(<4 x i32> %a0) {
;CHECK-LABEL: stack_fold_pmovsxdq
;CHECK: vpmovsxdq {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <4 x i64> @llvm.x86.avx2.pmovsxdq(<4 x i32> %a0)
+ %2 = sext <4 x i32> %a0 to <4 x i64>
ret <4 x i64> %2
}
-declare <4 x i64> @llvm.x86.avx2.pmovsxdq(<4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_pmovsxwd(<8 x i16> %a0) {
;CHECK-LABEL: stack_fold_pmovsxwd
;CHECK: vpmovsxwd {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{flags}"()
- %2 = call <8 x i32> @llvm.x86.avx2.pmovsxwd(<8 x i16> %a0)
+ %2 = sext <8 x i16> %a0 to <8 x i32>
ret <8 x i32> %2
}
-declare <8 x i32> @llvm.x86.avx2.pmovsxwd(<8 x i16>) nounwind readnone
define <4 x i64> @stack_fold_pmovsxwq(<8 x i16> %a0) {
;CHECK-LABEL: stack_fold_pmovsxwq