; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw,+avx512dq,+avx512vbmi,+avx512cd,+avx512vpopcntdq < %s | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-unknown"
; Stack reload folding tests.
;
; By including a nop call with sideeffects we can force a partial register spill of the
; relevant registers and check that the reload is correctly folded into the instruction.
define <8 x i32> @stack_fold_valignd_ymm(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: stack_fold_valignd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: valignd $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[1,2,3,4,5,6,7],ymm0[0]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
ret <8 x i32> %2
}
define <8 x i32> @stack_fold_valignd_ymm_mask(<8 x i32> %a, <8 x i32> %b, ptr %passthru, i8 %mask) {
; CHECK-LABEL: stack_fold_valignd_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: valignd $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm1 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm1 {%k1} = mem[1,2,3,4,5,6,7],ymm0[0]
; CHECK-NEXT: vmovdqa %ymm1, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
%3 = bitcast i8 %mask to <8 x i1>
%4 = load <8 x i32>, ptr %passthru
%5 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> %4
ret <8 x i32> %5
}
define <8 x i32> @stack_fold_valignd_ymm_maskz(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: stack_fold_valignd_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: valignd $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[1,2,3,4,5,6,7],ymm0[0]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> zeroinitializer
ret <8 x i32> %4
}
define <4 x i64> @stack_fold_valignq_ymm(<4 x i64> %a, <4 x i64> %b) {
; CHECK-LABEL: stack_fold_valignq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: valignq $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[1,2,3],ymm0[0]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 2, i32 3, i32 4>
ret <4 x i64> %2
}
define <16 x i8> @stack_fold_pavgb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pavgb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpavgb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <16 x i8> %a0 to <16 x i16>
%3 = zext <16 x i8> %a1 to <16 x i16>
%4 = add <16 x i16> %2, %3
%5 = add <16 x i16> %4, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
%6 = lshr <16 x i16> %5, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
%7 = trunc <16 x i16> %6 to <16 x i8>
ret <16 x i8> %7
}
define <32 x i8> @stack_fold_pavgb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pavgb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpavgb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <32 x i8> %a0 to <32 x i16>
%3 = zext <32 x i8> %a1 to <32 x i16>
%4 = add <32 x i16> %2, %3
%5 = add <32 x i16> %4, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
%6 = lshr <32 x i16> %5, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
%7 = trunc <32 x i16> %6 to <32 x i8>
ret <32 x i8> %7
}
define <8 x i16> @stack_fold_pavgw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pavgw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpavgw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <8 x i16> %a0 to <8 x i32>
%3 = zext <8 x i16> %a1 to <8 x i32>
%4 = add <8 x i32> %2, %3
%5 = add <8 x i32> %4, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
%6 = lshr <8 x i32> %5, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
%7 = trunc <8 x i32> %6 to <8 x i16>
ret <8 x i16> %7
}
define <16 x i16> @stack_fold_pavgw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pavgw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpavgw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <16 x i16> %a0 to <16 x i32>
%3 = zext <16 x i16> %a1 to <16 x i32>
%4 = add <16 x i32> %2, %3
%5 = add <16 x i32> %4, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
%6 = lshr <16 x i32> %5, <i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
%7 = trunc <16 x i32> %6 to <16 x i16>
ret <16 x i16> %7
}
define <4 x i32> @stack_fold_vpconflictd(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpconflictd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpconflictd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32> %a0)
ret <4 x i32> %2
}
define <8 x i32> @stack_fold_vpconflictd_ymm(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpconflictd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpconflictd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32> %a0)
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_vpconflictq(<2 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpconflictq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpconflictq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64> %a0)
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_vpconflictq_ymm(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpconflictq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpconflictq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64> %a0)
ret <4 x i64> %2
}
define <4 x i32> @stack_fold_extracti32x4(<8 x i16> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_extracti32x4:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovzxwd {{.*#+}} ymm0 = 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: vextracti128 $1, %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
; zext forces execution domain
%1 = zext <8 x i16> %a0 to <8 x i32>
%2 = shufflevector <8 x i32> %1, <8 x i32> %a1, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%3 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <4 x i32> %2
}
define <2 x i64> @stack_fold_extracti64x2(<4 x i32> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_extracti64x2:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; CHECK-NEXT: vextracti128 $1, %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
; zext forces execution domain
%1 = zext <4 x i32> %a0 to <4 x i64>
%2 = shufflevector <4 x i64> %1, <4 x i64> %a1, <2 x i32> <i32 2, i32 3>
%3 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <2 x i64> %2
}
define <8 x i32> @stack_fold_inserti32x4(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_inserti32x4:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vinserti128 $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubd %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> %a1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
; 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
}
define <4 x i64> @stack_fold_inserti64x2(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_inserti64x2:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vinserti128 $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubq %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <2 x i64> %a0, <2 x i64> %a1, <4 x i32> <i32 0, i32 1, 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
}
define <16 x i8> @stack_fold_pabsb(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pabsb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <16 x i8> %a0, zeroinitializer
%3 = sub <16 x i8> zeroinitializer, %a0
%4 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %3
ret <16 x i8> %4
}
define <32 x i8> @stack_fold_pabsb_ymm(<32 x i8> %a0) {
; CHECK-LABEL: stack_fold_pabsb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <32 x i8> %a0, zeroinitializer
%3 = sub <32 x i8> zeroinitializer, %a0
%4 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %3
ret <32 x i8> %4
}
define <4 x i32> @stack_fold_pabsd(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pabsd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <4 x i32> %a0, zeroinitializer
%3 = sub <4 x i32> zeroinitializer, %a0
%4 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %3
ret <4 x i32> %4
}
define <8 x i32> @stack_fold_pabsd_ymm(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_pabsd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <8 x i32> %a0, zeroinitializer
%3 = sub <8 x i32> zeroinitializer, %a0
%4 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %3
ret <8 x i32> %4
}
define <2 x i64> @stack_fold_pabsq(<2 x i64> %a0) {
; CHECK-LABEL: stack_fold_pabsq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <2 x i64> %a0, zeroinitializer
%3 = sub <2 x i64> zeroinitializer, %a0
%4 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %3
ret <2 x i64> %4
}
define <4 x i64> @stack_fold_pabsq_ymm(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_pabsq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <4 x i64> %a0, zeroinitializer
%3 = sub <4 x i64> zeroinitializer, %a0
%4 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %3
ret <4 x i64> %4
}
define <8 x i16> @stack_fold_pabsw(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pabsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <8 x i16> %a0, zeroinitializer
%3 = sub <8 x i16> zeroinitializer, %a0
%4 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %3
ret <8 x i16> %4
}
define <16 x i16> @stack_fold_pabsw_ymm(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_pabsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpabsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <16 x i16> %a0, zeroinitializer
%3 = sub <16 x i16> zeroinitializer, %a0
%4 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %3
ret <16 x i16> %4
}
define <8 x i16> @stack_fold_packssdw(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_packssdw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackssdw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32> %a0, <4 x i32> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.sse2.packssdw.128(<4 x i32>, <4 x i32>) nounwind readnone
define <16 x i16> @stack_fold_packssdw_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_packssdw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackssdw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.packssdw(<8 x i32> %a0, <8 x i32> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.packssdw(<8 x i32>, <8 x i32>) nounwind readnone
define <16 x i8> @stack_fold_packsswb(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_packsswb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpacksswb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.sse2.packsswb.128(<8 x i16> %a0, <8 x i16> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.x86.sse2.packsswb.128(<8 x i16>, <8 x i16>) nounwind readnone
define <32 x i8> @stack_fold_packsswb_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_packsswb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpacksswb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx2.packsswb(<16 x i16> %a0, <16 x i16> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.x86.avx2.packsswb(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_packusdw(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_packusdw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackusdw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.sse41.packusdw(<4 x i32> %a0, <4 x i32> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.sse41.packusdw(<4 x i32>, <4 x i32>) nounwind readnone
define <16 x i16> @stack_fold_packusdw_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_packusdw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackusdw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.packusdw(<8 x i32> %a0, <8 x i32> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.packusdw(<8 x i32>, <8 x i32>) nounwind readnone
define <16 x i8> @stack_fold_packuswb(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_packuswb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackuswb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16> %a0, <8 x i16> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.x86.sse2.packuswb.128(<8 x i16>, <8 x i16>) nounwind readnone
define <32 x i8> @stack_fold_packuswb_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_packuswb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpackuswb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx2.packuswb(<16 x i16> %a0, <16 x i16> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.x86.avx2.packuswb(<16 x i16>, <16 x i16>) nounwind readnone
define <16 x i8> @stack_fold_paddb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <16 x i8> %a0, %a1
ret <16 x i8> %2
}
define <16 x i8> @stack_fold_paddb_mask(<16 x i8> %a0, <16 x i8> %a1, ptr %a2, i16 %mask) {
; CHECK-LABEL: stack_fold_paddb_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <16 x i8> %a0, %a1
%3 = bitcast i16 %mask to <16 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <16 x i8>, ptr %a2
%5 = select <16 x i1> %3, <16 x i8> %2, <16 x i8> %4
ret <16 x i8> %5
}
define <16 x i8> @stack_fold_paddb_maskz(<16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_paddb_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <16 x i8> %a0, %a1
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i8> %2, <16 x i8> zeroinitializer
ret <16 x i8> %4
}
define <32 x i8> @stack_fold_paddb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <32 x i8> %a0, %a1
ret <32 x i8> %2
}
define <32 x i8> @stack_fold_paddb_mask_ymm(<32 x i8> %a0, <32 x i8> %a1, ptr %a2, i32 %mask) {
; CHECK-LABEL: stack_fold_paddb_mask_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <32 x i8> %a0, %a1
%3 = bitcast i32 %mask to <32 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <32 x i8>, ptr %a2
%5 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> %4
ret <32 x i8> %5
}
define <32 x i8> @stack_fold_paddb_maskz_ymm(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_paddb_maskz_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpaddb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <32 x i8> %a0, %a1
%3 = bitcast i32 %mask to <32 x i1>
%4 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> zeroinitializer
ret <32 x i8> %4
}
define <4 x i32> @stack_fold_paddd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_paddd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <4 x i32> %a0, %a1
ret <4 x i32> %2
}
define <8 x i32> @stack_fold_paddd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_paddd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <8 x i32> %a0, %a1
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_paddq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_paddq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <2 x i64> %a0, %a1
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_paddq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_paddq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <4 x i64> %a0, %a1
ret <4 x i64> %2
}
define <16 x i8> @stack_fold_paddsb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddsb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddsb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.sadd.sat.v16i8(<16 x i8> %a0, <16 x i8> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.sadd.sat.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
define <32 x i8> @stack_fold_paddsb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddsb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddsb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.sadd.sat.v32i8(<32 x i8> %a0, <32 x i8> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.sadd.sat.v32i8(<32 x i8>, <32 x i8>) nounwind readnone
define <8 x i16> @stack_fold_paddsw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_paddsw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.sadd.sat.v16i16(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.sadd.sat.v16i16(<16 x i16>, <16 x i16>) nounwind readnone
define <16 x i8> @stack_fold_paddusb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddusb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddusb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.uadd.sat.v16i8(<16 x i8> %a0, <16 x i8> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.uadd.sat.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
define <32 x i8> @stack_fold_paddusb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_paddusb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddusb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.uadd.sat.v32i8(<32 x i8> %a0, <32 x i8> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.uadd.sat.v32i8(<32 x i8>, <32 x i8>) nounwind readnone
define <8 x i16> @stack_fold_paddusw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddusw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddusw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_paddusw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddusw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddusw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_paddw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <8 x i16> %a0, %a1
ret <8 x i16> %2
}
define <16 x i16> @stack_fold_paddw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_paddw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpaddw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = add <16 x i16> %a0, %a1
ret <16 x i16> %2
}
define <32 x i8> @stack_fold_palignr(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_palignr:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpalignr $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0],mem[17,18,19,20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a1, <32 x i8> %a0, <32 x i32> <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, i32 32, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 48>
ret <32 x i8> %2
}
define <32 x i8> @stack_fold_palignr_mask(<32 x i8> %a0, <32 x i8> %a1, ptr %passthru, i32 %mask) {
; CHECK-LABEL: stack_fold_palignr_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpalignr $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm1 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm1 {%k1} = mem[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0],mem[17,18,19,20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16]
; CHECK-NEXT: vmovdqa %ymm1, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a1, <32 x i8> %a0, <32 x i32> <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, i32 32, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 48>
%3 = bitcast i32 %mask to <32 x i1>
%4 = load <32 x i8>, ptr %passthru
%5 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> %4
ret <32 x i8> %5
}
define <32 x i8> @stack_fold_palignr_maskz(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_palignr_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpalignr $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],ymm0[0],mem[17,18,19,20,21,22,23,24,25,26,27,28,29,30,31],ymm0[16]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a1, <32 x i8> %a0, <32 x i32> <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, i32 32, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 48>
%3 = bitcast i32 %mask to <32 x i1>
%4 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> zeroinitializer
ret <32 x i8> %4
}
define i16 @stack_fold_pcmpeqb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pcmpeqb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpcmpeqb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kmovd %k0, %eax
; CHECK-NEXT: # kill: def $ax killed $ax killed $eax
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp eq <16 x i8> %a0, %a1
%3 = bitcast <16 x i1> %2 to i16
ret i16 %3
}
define i8 @stack_fold_pcmpeqd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pcmpeqd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpcmpeqd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kmovd %k0, %eax
; CHECK-NEXT: # kill: def $al killed $al killed $eax
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp eq <4 x i32> %a0, %a1
%3 = shufflevector <4 x i1> %2, <4 x i1> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%4 = bitcast <8 x i1> %3 to i8
ret i8 %4
}
define i8 @stack_fold_pcmpeqq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_pcmpeqq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpcmpeqq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kmovd %k0, %eax
; CHECK-NEXT: # kill: def $al killed $al killed $eax
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp eq <2 x i64> %a0, %a1
%3 = shufflevector <2 x i1> %2, <2 x i1> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 2, i32 3, i32 2, i32 3>
%4 = bitcast <8 x i1> %3 to i8
ret i8 %4
}
define i8 @stack_fold_pcmpeqw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pcmpeqw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpcmpeqw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kmovd %k0, %eax
; CHECK-NEXT: # kill: def $al killed $al killed $eax
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp eq <8 x i16> %a0, %a1
%3 = bitcast <8 x i1> %2 to i8
ret i8 %3
}
define <32 x i8> @stack_fold_permbvar(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_permbvar:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpermb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubb %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx512.permvar.qi.256(<32 x i8> %a1, <32 x i8> %a0)
; add forces execution domain
%3 = add <32 x i8> %2, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
ret <32 x i8> %3
}
declare <32 x i8> @llvm.x86.avx512.permvar.qi.256(<32 x i8>, <32 x i8>) nounwind readonly
define <8 x i32> @stack_fold_permd(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_permd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubd %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a1, <8 x i32> %a0)
; 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.permd(<8 x i32>, <8 x i32>) nounwind readonly
define <16 x i8> @stack_fold_vpermi2b(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2) {
; CHECK-LABEL: stack_fold_vpermi2b:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2b {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.avx512.vpermi2var.qi.128(<16 x i8> %x1, <16 x i8> %x0, <16 x i8> %x2)
ret <16 x i8> %2
}
define <32 x i8> @stack_fold_vpermi2b_ymm(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2) {
; CHECK-LABEL: stack_fold_vpermi2b_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2b {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx512.vpermi2var.qi.256(<32 x i8> %x1, <32 x i8> %x0, <32 x i8> %x2)
ret <32 x i8> %2
}
define <4 x i32> @stack_fold_vpermi2d(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: stack_fold_vpermi2d:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2d {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx512.vpermi2var.d.128(<4 x i32> %x1, <4 x i32> %x0, <4 x i32> %x2)
ret <4 x i32> %2
}
define <8 x i32> @stack_fold_vpermi2d_ymm(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: stack_fold_vpermi2d_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2d {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx512.vpermi2var.d.256(<8 x i32> %x1, <8 x i32> %x0, <8 x i32> %x2)
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_vpermi2q(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: stack_fold_vpermi2q:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2q {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx512.vpermi2var.q.128(<2 x i64> %x1, <2 x i64> %x0, <2 x i64> %x2)
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_vpermi2q_ymm(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: stack_fold_vpermi2q_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2q {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.vpermi2var.q.256(<4 x i64> %x1, <4 x i64> %x0, <4 x i64> %x2)
ret <4 x i64> %2
}
define <8 x i16> @stack_fold_vpermi2w(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2) {
; CHECK-LABEL: stack_fold_vpermi2w:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2w {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.avx512.vpermi2var.hi.128(<8 x i16> %x1, <8 x i16> %x0, <8 x i16> %x2)
ret <8 x i16> %2
}
define <16 x i16> @stack_fold_vpermi2w_ymm(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2) {
; CHECK-LABEL: stack_fold_vpermi2w_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermi2w {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.vpermi2var.hi.256(<16 x i16> %x1, <16 x i16> %x0, <16 x i16> %x2)
ret <16 x i16> %2
}
define <4 x i64> @stack_fold_permq(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_permq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermq $235, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[3,2,2,3]
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubq %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%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
}
define <4 x i64> @stack_fold_permqvar(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_permqvar:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpermq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubq %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.permvar.di.256(<4 x i64> %a1, <4 x i64> %a0)
; 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.avx512.permvar.di.256(<4 x i64>, <4 x i64>) nounwind readonly
define <16 x i8> @stack_fold_vpermt2b(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2) {
; CHECK-LABEL: stack_fold_vpermt2b:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2b {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.avx512.vpermi2var.qi.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.x86.avx512.vpermi2var.qi.128(<16 x i8>, <16 x i8>, <16 x i8>)
define <32 x i8> @stack_fold_vpermt2b_ymm(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2) {
; CHECK-LABEL: stack_fold_vpermt2b_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2b {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx512.vpermi2var.qi.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.x86.avx512.vpermi2var.qi.256(<32 x i8>, <32 x i8>, <32 x i8>)
define <4 x i32> @stack_fold_vpermt2d(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2) {
; CHECK-LABEL: stack_fold_vpermt2d:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2d {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx512.vpermi2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.avx512.vpermi2var.d.128(<4 x i32>, <4 x i32>, <4 x i32>)
define <8 x i32> @stack_fold_vpermt2d_ymm(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2) {
; CHECK-LABEL: stack_fold_vpermt2d_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2d {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx512.vpermi2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx512.vpermi2var.d.256(<8 x i32>, <8 x i32>, <8 x i32>)
define <2 x i64> @stack_fold_vpermt2q(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2) {
; CHECK-LABEL: stack_fold_vpermt2q:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2q {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx512.vpermi2var.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.avx512.vpermi2var.q.128(<2 x i64>, <2 x i64>, <2 x i64>)
define <4 x i64> @stack_fold_vpermt2q_ymm(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2) {
; CHECK-LABEL: stack_fold_vpermt2q_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2q {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.vpermi2var.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx512.vpermi2var.q.256(<4 x i64>, <4 x i64>, <4 x i64>)
define <8 x i16> @stack_fold_vpermt2w(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2) {
; CHECK-LABEL: stack_fold_vpermt2w:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2w {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.avx512.vpermi2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.avx512.vpermi2var.hi.128(<8 x i16>, <8 x i16>, <8 x i16>)
define <16 x i16> @stack_fold_vpermt2w_ymm(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2) {
; CHECK-LABEL: stack_fold_vpermt2w_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpermt2w {{[-0-9]+}}(%r{{[sb]}}p), %ymm1, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.vpermi2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx512.vpermi2var.hi.256(<16 x i16>, <16 x i16>, <16 x i16>)
define <16 x i16> @stack_fold_permwvar(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_permwvar:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpermw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubw %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.permvar.hi.256(<16 x i16> %a1, <16 x i16> %a0)
; add forces execution domain
%3 = add <16 x i16> %2, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
ret <16 x i16> %3
}
declare <16 x i16> @llvm.x86.avx512.permvar.hi.256(<16 x i16>, <16 x i16>) nounwind readonly
define <4 x i32> @stack_fold_vplzcntd(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_vplzcntd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vplzcntd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %a0, i1 false)
ret <4 x i32> %2
}
define <8 x i32> @stack_fold_vplzcntd_ymm(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vplzcntd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vplzcntd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.ctlz.v8i32(<8 x i32> %a0, i1 false)
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_vplzcntq(<2 x i64> %a0) {
; CHECK-LABEL: stack_fold_vplzcntq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vplzcntq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %a0, i1 false)
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_vplzcntq_ymm(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vplzcntq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vplzcntq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.ctlz.v4i64(<4 x i64> %a0, i1 false)
ret <4 x i64> %2
}
define <8 x i16> @stack_fold_pmaddubsw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaddubsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8> %a0, <16 x i8> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8>, <16 x i8>) nounwind readnone
define <8 x i16> @stack_fold_pmaddubsw_mask(ptr %passthru, <16 x i8> %a0, <16 x i8> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddubsw_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8> %a0, <16 x i8> %a1)
%3 = bitcast i8 %mask to <8 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <8 x i16>, ptr %passthru
%5 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> %4
ret <8 x i16> %5
}
define <8 x i16> @stack_fold_pmaddubsw_maskz(<16 x i8> %a0, <16 x i8> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddubsw_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.ssse3.pmadd.ub.sw.128(<16 x i8> %a0, <16 x i8> %a1)
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> zeroinitializer
ret <8 x i16> %4
}
define <16 x i16> @stack_fold_pmaddubsw_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaddubsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8> %a0, <32 x i8> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8>, <32 x i8>) nounwind readnone
define <16 x i16> @stack_fold_pmaddubsw_ymm_mask(ptr %passthru, <32 x i8> %a0, <32 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_pmaddubsw_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8> %a0, <32 x i8> %a1)
%3 = bitcast i16 %mask to <16 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <16 x i16>, ptr %passthru
%5 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> %4
ret <16 x i16> %5
}
define <16 x i16> @stack_fold_pmaddubsw_ymm_maskz(<32 x i8> %a0, <32 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_pmaddubsw_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaddubsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.pmadd.ub.sw(<32 x i8> %a0, <32 x i8> %a1)
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> zeroinitializer
ret <16 x i16> %4
}
define <4 x i32> @stack_fold_pmaddwd(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaddwd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16> %a0, <8 x i16> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16>, <8 x i16>) nounwind readnone
define <4 x i32> @stack_fold_pmaddwd_mask(ptr %passthru, <8 x i16> %a0, <8 x i16> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddwd_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16> %a0, <8 x i16> %a1)
%3 = bitcast i8 %mask to <8 x i1>
%4 = shufflevector <8 x i1> %3, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
; load needed to keep the operation from being scheduled about the asm block
%5 = load <4 x i32>, ptr %passthru
%6 = select <4 x i1> %4, <4 x i32> %2, <4 x i32> %5
ret <4 x i32> %6
}
define <4 x i32> @stack_fold_pmaddwd_maskz(<8 x i16> %a0, <8 x i16> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddwd_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.pmadd.wd(<8 x i16> %a0, <8 x i16> %a1)
%3 = bitcast i8 %mask to <8 x i1>
%4 = shufflevector <8 x i1> %3, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%5 = select <4 x i1> %4, <4 x i32> %2, <4 x i32> zeroinitializer
ret <4 x i32> %5
}
define <8 x i32> @stack_fold_pmaddwd_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaddwd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16> %a0, <16 x i16> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i32> @stack_fold_pmaddwd_ymm_mask(ptr %passthru, <16 x i16> %a0, <16 x i16> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddwd_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16> %a0, <16 x i16> %a1)
%3 = bitcast i8 %mask to <8 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <8 x i32>, ptr %passthru
%5 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> %4
ret <8 x i32> %5
}
define <8 x i32> @stack_fold_pmaddwd_ymm_maskz(<16 x i16> %a0, <16 x i16> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaddwd_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaddwd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.pmadd.wd(<16 x i16> %a0, <16 x i16> %a1)
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> zeroinitializer
ret <8 x i32> %4
}
define <16 x i8> @stack_fold_pmaxsb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaxsb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <16 x i8> %a0, %a1
%3 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %a1
ret <16 x i8> %3
}
define <32 x i8> @stack_fold_pmaxsb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaxsb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <32 x i8> %a0, %a1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %a1
ret <32 x i8> %3
}
define <4 x i32> @stack_fold_pmaxsd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmaxsd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <4 x i32> %a0, %a1
%3 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %a1
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmaxsd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmaxsd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <8 x i32> %a0, %a1
%3 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %a1
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pmaxsq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_pmaxsq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <2 x i64> %a0, %a1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %a1
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmaxsq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_pmaxsq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <4 x i64> %a0, %a1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %a1
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_pmaxsw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaxsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <8 x i16> %a0, %a1
%3 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %a1
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pmaxsw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaxsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp sgt <16 x i16> %a0, %a1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %a1
ret <16 x i16> %3
}
define <16 x i8> @stack_fold_pmaxub(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaxub:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxub {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <16 x i8> %a0, %a1
%3 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %a1
ret <16 x i8> %3
}
define <32 x i8> @stack_fold_pmaxub_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pmaxub_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxub {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <32 x i8> %a0, %a1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %a1
ret <32 x i8> %3
}
define <4 x i32> @stack_fold_pmaxud(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmaxud:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxud {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <4 x i32> %a0, %a1
%3 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %a1
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmaxud_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmaxud_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxud {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <8 x i32> %a0, %a1
%3 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %a1
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pmaxuq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_pmaxuq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <2 x i64> %a0, %a1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %a1
ret <2 x i64> %3
}
define <2 x i64> @stack_fold_pmaxuq_mask(ptr %passthru, <2 x i64> %a0, <2 x i64> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaxuq_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <2 x i64>, ptr %passthru
%3 = icmp ugt <2 x i64> %a0, %a1
%4 = select <2 x i1> %3, <2 x i64> %a0, <2 x i64> %a1
%5 = bitcast i8 %mask to <8 x i1>
%extract = shufflevector <8 x i1> %5, <8 x i1> %5, <2 x i32> <i32 0, i32 1>
%6 = select <2 x i1> %extract, <2 x i64> %4, <2 x i64> %2
ret <2 x i64> %6
}
define <2 x i64> @stack_fold_pmaxuq_maskz(<2 x i64> %a0, <2 x i64> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaxuq_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <2 x i64> %a0, %a1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %a1
%4 = bitcast i8 %mask to <8 x i1>
%extract = shufflevector <8 x i1> %4, <8 x i1> %4, <2 x i32> <i32 0, i32 1>
%5 = select <2 x i1> %extract, <2 x i64> %3, <2 x i64> zeroinitializer
ret <2 x i64> %5
}
define <4 x i64> @stack_fold_pmaxuq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_pmaxuq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <4 x i64> %a0, %a1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %a1
ret <4 x i64> %3
}
define <4 x i64> @stack_fold_pmaxuq_ymm_mask(ptr %passthru, <4 x i64> %a0, <4 x i64> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaxuq_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <4 x i64>, ptr %passthru
%3 = icmp ugt <4 x i64> %a0, %a1
%4 = select <4 x i1> %3, <4 x i64> %a0, <4 x i64> %a1
%5 = bitcast i8 %mask to <8 x i1>
%extract = shufflevector <8 x i1> %5, <8 x i1> %5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %extract, <4 x i64> %4, <4 x i64> %2
ret <4 x i64> %6
}
define <4 x i64> @stack_fold_pmaxuq_ymm_maskz(<4 x i64> %a0, <4 x i64> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmaxuq_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmaxuq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <4 x i64> %a0, %a1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %a1
%4 = bitcast i8 %mask to <8 x i1>
%extract = shufflevector <8 x i1> %4, <8 x i1> %4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%5 = select <4 x i1> %extract, <4 x i64> %3, <4 x i64> zeroinitializer
ret <4 x i64> %5
}
define <8 x i16> @stack_fold_pmaxuw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaxuw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxuw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <8 x i16> %a0, %a1
%3 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %a1
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pmaxuw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pmaxuw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmaxuw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ugt <16 x i16> %a0, %a1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %a1
ret <16 x i16> %3
}
define <16 x i8> @stack_fold_pminsb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pminsb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <16 x i8> %a0, %a1
%3 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %a1
ret <16 x i8> %3
}
define <32 x i8> @stack_fold_pminsb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pminsb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <32 x i8> %a0, %a1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %a1
ret <32 x i8> %3
}
define <4 x i32> @stack_fold_pminsd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pminsd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <4 x i32> %a0, %a1
%3 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %a1
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pminsd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pminsd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <8 x i32> %a0, %a1
%3 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %a1
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pminsq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_pminsq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <2 x i64> %a0, %a1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %a1
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pminsq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_pminsq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <4 x i64> %a0, %a1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %a1
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_pminsw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pminsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <8 x i16> %a0, %a1
%3 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %a1
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pminsw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pminsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp slt <16 x i16> %a0, %a1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %a1
ret <16 x i16> %3
}
define <16 x i8> @stack_fold_pminub(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pminub:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminub {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <16 x i8> %a0, %a1
%3 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %a1
ret <16 x i8> %3
}
define <32 x i8> @stack_fold_pminub_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pminub_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminub {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <32 x i8> %a0, %a1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %a1
ret <32 x i8> %3
}
define <4 x i32> @stack_fold_pminud(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pminud:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminud {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <4 x i32> %a0, %a1
%3 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %a1
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pminud_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pminud_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminud {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <8 x i32> %a0, %a1
%3 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %a1
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pminuq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_pminuq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminuq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <2 x i64> %a0, %a1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %a1
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pminuq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_pminuq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminuq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <4 x i64> %a0, %a1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %a1
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_pminuw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_pminuw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminuw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <8 x i16> %a0, %a1
%3 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %a1
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pminuw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_pminuw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpminuw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = icmp ult <16 x i16> %a0, %a1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %a1
ret <16 x i16> %3
}
define <8 x i16> @stack_fold_vpmovdw(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpmovdw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovdw %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32> %a0, <8 x i16> undef, i8 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <8 x i16> %1
}
declare <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32>, <8 x i16>, i8)
define <4 x i32> @stack_fold_vpmovqd(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpmovqd:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovqd %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = trunc <4 x i64> %a0 to <4 x i32>
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <4 x i32> %1
}
declare <4 x i32> @llvm.x86.avx512.mask.pmov.qd.256(<4 x i64>, <4 x i32>, i8)
define <16 x i8> @stack_fold_vpmovwb(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_vpmovwb:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovwb %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = trunc <16 x i16> %a0 to <16 x i8>
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <16 x i8> %1
}
declare <16 x i8> @llvm.x86.avx512.mask.pmov.wb.256(<16 x i16>, <16 x i8>, i16)
define <8 x i16> @stack_fold_vpmovsdw(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpmovsdw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovsdw %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32> %a0, <8 x i16> undef, i8 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <8 x i16> %1
}
declare <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32>, <8 x i16>, i8)
define <4 x i32> @stack_fold_vpmovsqd(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpmovsqd:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovsqd %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64> %a0, <4 x i32> undef, i8 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <4 x i32> %1
}
declare <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64>, <4 x i32>, i8)
define <16 x i8> @stack_fold_vpmovswb(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_vpmovswb:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovswb %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.256(<16 x i16> %a0, <16 x i8> undef, i16 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <16 x i8> %1
}
declare <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.256(<16 x i16>, <16 x i8>, i16)
define <4 x i32> @stack_fold_pmovsxbd(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = sext <4 x i8> %2 to <4 x i32>
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmovsxbd_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%3 = sext <8 x i8> %2 to <8 x i32>
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pmovsxbq(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <2 x i32> <i32 0, i32 1>
%3 = sext <2 x i8> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovsxbq_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = sext <4 x i8> %2 to <4 x i64>
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_pmovsxbw(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%3 = sext <8 x i8> %2 to <8 x i16>
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pmovsxbw_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovsxbw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sext <16 x i8> %a0 to <16 x i16>
ret <16 x i16> %2
}
define <2 x i64> @stack_fold_pmovsxdq(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pmovsxdq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxdq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%3 = sext <2 x i32> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovsxdq_ymm(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pmovsxdq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxdq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sext <4 x i32> %a0 to <4 x i64>
ret <4 x i64> %2
}
define <4 x i32> @stack_fold_pmovsxwd(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovsxwd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxwd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = sext <4 x i16> %2 to <4 x i32>
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmovsxwd_ymm(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovsxwd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxwd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sext <8 x i16> %a0 to <8 x i32>
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_pmovsxwq(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovsxwq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxwq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <2 x i32> <i32 0, i32 1>
%3 = sext <2 x i16> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovsxwq_ymm(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovsxwq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovsxwq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = sext <4 x i16> %2 to <4 x i64>
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_vpmovusdw(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpmovusdw:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovusdw %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32> %a0, <8 x i16> undef, i8 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <8 x i16> %1
}
declare <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32>, <8 x i16>, i8)
define <4 x i32> @stack_fold_vpmovusqd(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpmovusqd:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovusqd %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64> %a0, <4 x i32> undef, i8 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <4 x i32> %1
}
declare <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64>, <4 x i32>, i8)
define <16 x i8> @stack_fold_vpmovuswb(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_vpmovuswb:
; CHECK: # %bb.0:
; CHECK-NEXT: vpmovuswb %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Folded Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.256(<16 x i16> %a0, <16 x i8> undef, i16 -1)
%2 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
ret <16 x i8> %1
}
declare <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.256(<16 x i16>, <16 x i8>, i16)
define <4 x i32> @stack_fold_pmovzxbd(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 17, i32 18, i32 1, i32 19, i32 20, i32 21, i32 2, i32 22, i32 23, i32 24, i32 3, i32 25, i32 26, i32 27>
%3 = bitcast <16 x i8> %2 to <4 x i32>
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmovzxbd_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero,mem[4],zero,zero,zero,mem[5],zero,zero,zero,mem[6],zero,zero,zero,mem[7],zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%3 = zext <8 x i8> %2 to <8 x i32>
ret <8 x i32> %3
}
define <2 x i64> @stack_fold_pmovzxbq(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,zero,zero,zero,zero,zero,zero,mem[1],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 1, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28>
%3 = bitcast <16 x i8> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovzxbq_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,zero,zero,zero,zero,zero,zero,mem[1],zero,zero,zero,zero,zero,zero,zero,mem[2],zero,zero,zero,zero,zero,zero,zero,mem[3],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = zext <4 x i8> %2 to <4 x i64>
ret <4 x i64> %3
}
define <8 x i16> @stack_fold_pmovzxbw(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> zeroinitializer, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
%3 = bitcast <16 x i8> %2 to <8 x i16>
ret <8 x i16> %3
}
define <16 x i16> @stack_fold_pmovzxbw_ymm(<16 x i8> %a0) {
; CHECK-LABEL: stack_fold_pmovzxbw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero,mem[8],zero,mem[9],zero,mem[10],zero,mem[11],zero,mem[12],zero,mem[13],zero,mem[14],zero,mem[15],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <16 x i8> %a0 to <16 x i16>
ret <16 x i16> %2
}
define <2 x i64> @stack_fold_pmovzxdq(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pmovzxdq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxdq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,mem[1],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> zeroinitializer, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
%3 = bitcast <4 x i32> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovzxdq_ymm(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pmovzxdq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxdq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <4 x i32> %a0 to <4 x i64>
ret <4 x i64> %2
}
define <4 x i32> @stack_fold_pmovzxwd(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovzxwd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxwd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> zeroinitializer, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
%3 = bitcast <8 x i16> %2 to <4 x i32>
ret <4 x i32> %3
}
define <8 x i32> @stack_fold_pmovzxwd_ymm(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovzxwd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxwd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,mem[1],zero,mem[2],zero,mem[3],zero,mem[4],zero,mem[5],zero,mem[6],zero,mem[7],zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = zext <8 x i16> %a0 to <8 x i32>
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_pmovzxwq(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovzxwq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxwq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> zeroinitializer, <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 1, i32 11, i32 12, i32 13>
%3 = bitcast <8 x i16> %2 to <2 x i64>
ret <2 x i64> %3
}
define <4 x i64> @stack_fold_pmovzxwq_ymm(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pmovzxwq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmovzxwq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = zext <4 x i16> %2 to <4 x i64>
ret <4 x i64> %3
}
define <4 x i64> @stack_fold_pmovzxwq_maskz_ymm(<8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pmovzxwq_maskz_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmovzxwq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = zext <4 x i16> %2 to <4 x i64>
%4 = bitcast i8 %mask to <8 x i1>
%5 = shufflevector <8 x i1> %4, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %5, <4 x i64> %3, <4 x i64> zeroinitializer
ret <4 x i64> %6
}
define <4 x i64> @stack_fold_pmovzxwq_mask_ymm(<4 x i64> %passthru, <8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pmovzxwq_mask_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpmovzxwq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%3 = zext <4 x i16> %2 to <4 x i64>
%4 = bitcast i8 %mask to <8 x i1>
%5 = shufflevector <8 x i1> %4, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%6 = select <4 x i1> %5, <4 x i64> %3, <4 x i64> %passthru
ret <4 x i64> %6
}
define <2 x i64> @stack_fold_pmuldq(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmuldq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmuldq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <4 x i32> %a0 to <2 x i64>
%3 = bitcast <4 x i32> %a1 to <2 x i64>
%4 = shl <2 x i64> %2, <i64 32, i64 32>
%5 = ashr <2 x i64> %4, <i64 32, i64 32>
%6 = shl <2 x i64> %3, <i64 32, i64 32>
%7 = ashr <2 x i64> %6, <i64 32, i64 32>
%8 = mul <2 x i64> %5, %7
ret <2 x i64> %8
}
define <4 x i64> @stack_fold_pmuldq_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmuldq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmuldq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <8 x i32> %a0 to <4 x i64>
%3 = bitcast <8 x i32> %a1 to <4 x i64>
%4 = shl <4 x i64> %2, <i64 32, i64 32, i64 32, i64 32>
%5 = ashr <4 x i64> %4, <i64 32, i64 32, i64 32, i64 32>
%6 = shl <4 x i64> %3, <i64 32, i64 32, i64 32, i64 32>
%7 = ashr <4 x i64> %6, <i64 32, i64 32, i64 32, i64 32>
%8 = mul <4 x i64> %5, %7
ret <4 x i64> %8
}
define <2 x i64> @stack_fold_pmuludq(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmuludq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmuludq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <4 x i32> %a0 to <2 x i64>
%3 = bitcast <4 x i32> %a1 to <2 x i64>
%4 = and <2 x i64> %2, <i64 4294967295, i64 4294967295>
%5 = and <2 x i64> %3, <i64 4294967295, i64 4294967295>
%6 = mul <2 x i64> %4, %5
ret <2 x i64> %6
}
define <4 x i64> @stack_fold_pmuludq_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_pmuludq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpmuludq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <8 x i32> %a0 to <4 x i64>
%3 = bitcast <8 x i32> %a1 to <4 x i64>
%4 = and <4 x i64> %2, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%5 = and <4 x i64> %3, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%6 = mul <4 x i64> %4, %5
ret <4 x i64> %6
}
define <4 x i64> @stack_fold_pmuludq_ymm_mask(ptr %passthru, <8 x i32> %a0, <8 x i32> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmuludq_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpmuludq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm1 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm1, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <8 x i32> %a0 to <4 x i64>
%3 = bitcast <8 x i32> %a1 to <4 x i64>
%4 = and <4 x i64> %2, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%5 = and <4 x i64> %3, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%6 = mul <4 x i64> %4, %5
%7 = bitcast i8 %mask to <8 x i1>
%8 = shufflevector <8 x i1> %7, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = load <4 x i64>, ptr %passthru
%10 = select <4 x i1> %8, <4 x i64> %6, <4 x i64> %9
ret <4 x i64> %10
}
define <4 x i64> @stack_fold_pmuludq_ymm_maskz(<8 x i32> %a0, <8 x i32> %a1, i8 %mask) {
; CHECK-LABEL: stack_fold_pmuludq_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpmuludq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = bitcast <8 x i32> %a0 to <4 x i64>
%3 = bitcast <8 x i32> %a1 to <4 x i64>
%4 = and <4 x i64> %2, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%5 = and <4 x i64> %3, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%6 = mul <4 x i64> %4, %5
%7 = bitcast i8 %mask to <8 x i1>
%8 = shufflevector <8 x i1> %7, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%9 = select <4 x i1> %8, <4 x i64> %6, <4 x i64> zeroinitializer
ret <4 x i64> %9
}
define <4 x i32> @stack_fold_vpopcntd(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpopcntd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpopcntd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> %a0)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.ctpop.v4i32(<4 x i32>) nounwind readonly
define <8 x i32> @stack_fold_vpopcntd_ymm(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_vpopcntd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpopcntd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.ctpop.v8i32(<8 x i32> %a0)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.ctpop.v8i32(<8 x i32>) nounwind readonly
define <2 x i64> @stack_fold_vpopcntq(<2 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpopcntq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpopcntq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %a0)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.ctpop.v2i64(<2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_vpopcntq_ymm(<4 x i64> %a0) {
; CHECK-LABEL: stack_fold_vpopcntq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpopcntq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.ctpop.v4i64(<4 x i64> %a0)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.ctpop.v4i64(<4 x i64>) nounwind readnone
define <2 x i64> @stack_fold_psadbw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_psadbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsadbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.sse2.psad.bw(<16 x i8> %a0, <16 x i8> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.sse2.psad.bw(<16 x i8>, <16 x i8>) nounwind readnone
define <2 x i64> @stack_fold_psadbw_commute(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_psadbw_commute:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsadbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.sse2.psad.bw(<16 x i8> %a1, <16 x i8> %a0)
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_psadbw_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_psadbw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsadbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psad.bw(<32 x i8> %a0, <32 x i8> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx2.psad.bw(<32 x i8>, <32 x i8>) nounwind readnone
define <4 x i64> @stack_fold_psadbw_ymm_commute(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_psadbw_ymm_commute:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsadbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psad.bw(<32 x i8> %a1, <32 x i8> %a0)
ret <4 x i64> %2
}
define <16 x i8> @stack_fold_pshufb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_pshufb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a0, <16 x i8> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8>, <16 x i8>) nounwind readnone
define <16 x i8> @stack_fold_pshufb_mask(ptr %passthru, <16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_pshufb_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <16 x i8>, ptr %passthru
%3 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a0, <16 x i8> %a1)
%4 = bitcast i16 %mask to <16 x i1>
%5 = select <16 x i1> %4, <16 x i8> %3, <16 x i8> %2
ret <16 x i8> %5
}
define <16 x i8> @stack_fold_pshufb_maskz(<16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_pshufb_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a0, <16 x i8> %a1)
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i8> %2, <16 x i8> zeroinitializer
ret <16 x i8> %4
}
define <32 x i8> @stack_fold_pshufb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_pshufb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8>, <32 x i8>)
define <32 x i8> @stack_fold_pshufb_ymm_mask(ptr %passthru, <32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_pshufb_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <32 x i8>, ptr %passthru
%3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> %a1)
%4 = bitcast i32 %mask to <32 x i1>
%5 = select <32 x i1> %4, <32 x i8> %3, <32 x i8> %2
ret <32 x i8> %5
}
define <32 x i8> @stack_fold_pshufb_ymm_maskz(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_pshufb_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> %a1)
%3 = bitcast i32 %mask to <32 x i1>
%4 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> zeroinitializer
ret <32 x i8> %4
}
define <4 x i32> @stack_fold_pshufd(<4 x i32> %a0) {
; CHECK-LABEL: stack_fold_pshufd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[3,2,1,0]
; CHECK-NEXT: vpcmpeqd %xmm1, %xmm1, %xmm1
; CHECK-NEXT: vpsubd %xmm1, %xmm0, %xmm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
%3 = add <4 x i32> %2, <i32 1, i32 1, i32 1, i32 1>
ret <4 x i32> %3
}
define <4 x i32> @stack_fold_pshufd_mask(<4 x i32> %passthru, <4 x i32> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufd_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} = mem[3,2,1,0]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
%3 = bitcast i8 %mask to <8 x i1>
%4 = shufflevector <8 x i1> %3, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%5 = select <4 x i1> %4, <4 x i32> %2, <4 x i32> %passthru
ret <4 x i32> %5
}
define <4 x i32> @stack_fold_pshufd_maskz(<4 x i32> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufd_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} {z} = mem[3,2,1,0]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i32> %a0, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
%3 = bitcast i8 %mask to <8 x i1>
%4 = shufflevector <8 x i1> %3, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%5 = select <4 x i1> %4, <4 x i32> %2, <4 x i32> zeroinitializer
ret <4 x i32> %5
}
define <8 x i32> @stack_fold_pshufd_ymm(<8 x i32> %a0) {
; CHECK-LABEL: stack_fold_pshufd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[3,2,1,0,7,6,5,4]
; CHECK-NEXT: vpcmpeqd %ymm1, %ymm1, %ymm1
; CHECK-NEXT: vpsubd %ymm1, %ymm0, %ymm0
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
%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
}
define <8 x i32> @stack_fold_pshufd_ymm_mask(<8 x i32> %passthru, <8 x i32> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufd_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} = mem[3,2,1,0,7,6,5,4]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> %passthru
ret <8 x i32> %4
}
define <8 x i32> @stack_fold_pshufd_ymm_maskz(<8 x i32> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufd_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufd $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[3,2,1,0,7,6,5,4]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> zeroinitializer
ret <8 x i32> %4
}
define <8 x i16> @stack_fold_pshufhw(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pshufhw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufhw $11, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[0,1,2,3,7,6,4,4]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 4, i32 4>
ret <8 x i16> %2
}
define <8 x i16> @stack_fold_pshufhw_mask(<8 x i16> %passthru, <8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufhw_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vpshufhw $11, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} = mem[0,1,2,3,7,6,4,4]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 4, i32 4>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> %passthru
ret <8 x i16> %4
}
define <8 x i16> @stack_fold_pshufhw_maskz(<8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshufhw_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufhw $11, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} {z} = mem[0,1,2,3,7,6,4,4]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 4, i32 4>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> zeroinitializer
ret <8 x i16> %4
}
define <16 x i16> @stack_fold_pshufhw_ymm(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_pshufhw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshufhw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[0,1,2,3,7,6,5,4,8,9,10,11,15,14,13,12]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 5, i32 4, i32 8, i32 9, i32 10, i32 11, i32 15, i32 14, i32 13, i32 12>
ret <16 x i16> %2
}
define <16 x i16> @stack_fold_pshufhw_ymm_mask(<16 x i16> %passthru, <16 x i16> %a0, i16 %mask) {
; CHECK-LABEL: stack_fold_pshufhw_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpshufhw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} = mem[0,1,2,3,7,6,5,4,8,9,10,11,15,14,13,12]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 5, i32 4, i32 8, i32 9, i32 10, i32 11, i32 15, i32 14, i32 13, i32 12>
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> %passthru
ret <16 x i16> %4
}
define <16 x i16> @stack_fold_pshufhw_ymm_maskz(<16 x i16> %a0, i16 %mask) {
; CHECK-LABEL: stack_fold_pshufhw_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshufhw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[0,1,2,3,7,6,5,4,8,9,10,11,15,14,13,12]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 7, i32 6, i32 5, i32 4, i32 8, i32 9, i32 10, i32 11, i32 15, i32 14, i32 13, i32 12>
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> zeroinitializer
ret <16 x i16> %4
}
define <8 x i16> @stack_fold_pshuflw(<8 x i16> %a0) {
; CHECK-LABEL: stack_fold_pshuflw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[3,2,1,0,4,5,6,7]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
ret <8 x i16> %2
}
define <8 x i16> @stack_fold_pshuflw_mask(<8 x i16> %passthru, <8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshuflw_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} = mem[3,2,1,0,4,5,6,7]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> %passthru
ret <8 x i16> %4
}
define <8 x i16> @stack_fold_pshuflw_maskz(<8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: stack_fold_pshuflw_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} {z} = mem[3,2,1,0,4,5,6,7]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i16> %2, <8 x i16> zeroinitializer
ret <8 x i16> %4
}
define <16 x i16> @stack_fold_pshuflw_ymm(<16 x i16> %a0) {
; CHECK-LABEL: stack_fold_pshuflw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[3,2,1,0,4,5,6,7,11,10,9,8,12,13,14,15]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7, i32 11, i32 10, i32 9, i32 8, i32 12, i32 13, i32 14, i32 15>
ret <16 x i16> %2
}
define <16 x i16> @stack_fold_pshuflw_ymm_mask(<16 x i16> %passthru, <16 x i16> %a0, i16 %mask) {
; CHECK-LABEL: stack_fold_pshuflw_ymm_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} = mem[3,2,1,0,4,5,6,7,11,10,9,8,12,13,14,15]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7, i32 11, i32 10, i32 9, i32 8, i32 12, i32 13, i32 14, i32 15>
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> %passthru
ret <16 x i16> %4
}
define <16 x i16> @stack_fold_pshuflw_ymm_maskz(<16 x i16> %a0, i16 %mask) {
; CHECK-LABEL: stack_fold_pshuflw_ymm_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpshuflw $27, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = mem[3,2,1,0,4,5,6,7,11,10,9,8,12,13,14,15]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7, i32 11, i32 10, i32 9, i32 8, i32 12, i32 13, i32 14, i32 15>
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i16> %2, <16 x i16> zeroinitializer
ret <16 x i16> %4
}
define <4 x i32> @stack_fold_pslld(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pslld:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpslld {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.psll.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.sse2.psll.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_pslld_ymm(<8 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_pslld_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpslld {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psll.d(<8 x i32> %a0, <4 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psll.d(<8 x i32>, <4 x i32>) nounwind readnone
define <16 x i8> @stack_fold_pslldq(<16 x i8> %a) {
; CHECK-LABEL: stack_fold_pslldq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpslldq $12, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,mem[0,1,2,3]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 16, i32 17, i32 18, i32 19>
ret <16 x i8> %2
}
define <32 x i8> @stack_fold_pslldq_ymm(<32 x i8> %a) {
; CHECK-LABEL: stack_fold_pslldq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpslldq $15, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,mem[0],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,mem[16]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> zeroinitializer, <32 x i8> %a, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 32, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 48>
ret <32 x i8> %2
}
define <2 x i64> @stack_fold_psllq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psllq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.sse2.psll.q(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.sse2.psll.q(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psllq_ymm(<4 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psllq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psll.q(<4 x i64> %a0, <2 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx2.psll.q(<4 x i64>, <2 x i64>) nounwind readnone
define <4 x i32> @stack_fold_psllvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psllvd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.avx2.psllv.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_psllvd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_psllvd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32> %a0, <8 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psllv.d.256(<8 x i32>, <8 x i32>) nounwind readnone
define <2 x i64> @stack_fold_psllvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psllvq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.avx2.psllv.q(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psllvq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_psllvq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64> %a0, <4 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx2.psllv.q.256(<4 x i64>, <4 x i64>) nounwind readnone
define <8 x i16> @stack_fold_psllvw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psllvw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.avx512.psllv.w.128(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psllvw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psllvw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllvw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx512.psllv.w.256(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_psllw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psllw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.sse2.psll.w(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.sse2.psll.w(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psllw_ymm(<16 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psllw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsllw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.psll.w(<16 x i16> %a0, <8 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.psll.w(<16 x i16>, <8 x i16>) nounwind readnone
define <4 x i32> @stack_fold_psrad(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrad:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrad {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.psra.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.sse2.psra.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_psrad_ymm(<8 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrad_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrad {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psra.d(<8 x i32> %a0, <4 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psra.d(<8 x i32>, <4 x i32>) nounwind readnone
define <2 x i64> @stack_fold_psraq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psraq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsraq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx512.psra.q.128(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.avx512.psra.q.128(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psraq_ymm(<4 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psraq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsraq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <4 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.psra.q.256(<4 x i64> %a0, <2 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx512.psra.q.256(<4 x i64>, <2 x i64>) nounwind readnone
define <4 x i32> @stack_fold_psravd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psravd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.avx2.psrav.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_psravd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_psravd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32> %a0, <8 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psrav.d.256(<8 x i32>, <8 x i32>) nounwind readnone
define <2 x i64> @stack_fold_psravq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psravq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx512.psrav.q.128(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.avx512.psrav.q.128(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psravq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_psravq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx512.psrav.q.256(<4 x i64> %a0, <4 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx512.psrav.q.256(<4 x i64>, <4 x i64>) nounwind readnone
define <8 x i16> @stack_fold_psravw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psravw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.avx512.psrav.w.128(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psravw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psravw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsravw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx512.psrav.w.256(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_psraw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psraw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsraw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.sse2.psra.w(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.sse2.psra.w(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psraw_ymm(<16 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psraw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsraw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.psra.w(<16 x i16> %a0, <8 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.psra.w(<16 x i16>, <8 x i16>) nounwind readnone
define <4 x i32> @stack_fold_psrld(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrld:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrld {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.sse2.psrl.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.sse2.psrl.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_psrld_ymm(<8 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrld_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vpsrld {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psrl.d(<8 x i32> %a0, <4 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psrl.d(<8 x i32>, <4 x i32>) nounwind readnone
define <16 x i8> @stack_fold_psrldq(<16 x i8> %a) {
; CHECK-LABEL: stack_fold_psrldq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrldq $12, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = mem[12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> zeroinitializer, <16 x i8> %a, <16 x i32> <i32 28, i32 29, i32 30, i32 31, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 09, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <16 x i8> %2
}
define <32 x i8> @stack_fold_psrldq_ymm(<32 x i8> %a) {
; CHECK-LABEL: stack_fold_psrldq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrldq $15, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = mem[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,mem[31],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> zeroinitializer, <32 x i8> %a, <32 x i32> <i32 47, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 63, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %2
}
define <2 x i64> @stack_fold_psrlq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psrlq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.sse2.psrl.q(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.sse2.psrl.q(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psrlq_ymm(<4 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psrlq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psrl.q(<4 x i64> %a0, <2 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx2.psrl.q(<4 x i64>, <2 x i64>) nounwind readnone
define <4 x i32> @stack_fold_psrlvd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrlvd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32> %a0, <4 x i32> %a1)
ret <4 x i32> %2
}
declare <4 x i32> @llvm.x86.avx2.psrlv.d(<4 x i32>, <4 x i32>) nounwind readnone
define <8 x i32> @stack_fold_psrlvd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_psrlvd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32> %a0, <8 x i32> %a1)
ret <8 x i32> %2
}
declare <8 x i32> @llvm.x86.avx2.psrlv.d.256(<8 x i32>, <8 x i32>) nounwind readnone
define <2 x i64> @stack_fold_psrlvq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psrlvq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64> %a0, <2 x i64> %a1)
ret <2 x i64> %2
}
declare <2 x i64> @llvm.x86.avx2.psrlv.q(<2 x i64>, <2 x i64>) nounwind readnone
define <4 x i64> @stack_fold_psrlvq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_psrlvq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64> %a0, <4 x i64> %a1)
ret <4 x i64> %2
}
declare <4 x i64> @llvm.x86.avx2.psrlv.q.256(<4 x i64>, <4 x i64>) nounwind readnone
define <8 x i16> @stack_fold_psrlvw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psrlvw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.avx512.psrlv.w.128(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.avx512.psrlv.w.128(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psrlvw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psrlvw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlvw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx512.psrlv.w.256(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx512.psrlv.w.256(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_psrlw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psrlw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.x86.sse2.psrl.w(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.x86.sse2.psrl.w(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psrlw_ymm(<16 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psrlw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsrlw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.x86.avx2.psrl.w(<16 x i16> %a0, <8 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.x86.avx2.psrl.w(<16 x i16>, <8 x i16>) nounwind readnone
define <16 x i8> @stack_fold_psubb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <16 x i8> %a0, %a1
ret <16 x i8> %2
}
define <32 x i8> @stack_fold_psubb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <32 x i8> %a0, %a1
ret <32 x i8> %2
}
define <4 x i32> @stack_fold_psubd(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: stack_fold_psubd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <4 x i32> %a0, %a1
ret <4 x i32> %2
}
define <8 x i32> @stack_fold_psubd_ymm(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: stack_fold_psubd_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <8 x i32> %a0, %a1
ret <8 x i32> %2
}
define <2 x i64> @stack_fold_psubq(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: stack_fold_psubq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <2 x i64> %a0, %a1
ret <2 x i64> %2
}
define <4 x i64> @stack_fold_psubq_ymm(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: stack_fold_psubq_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <4 x i64> %a0, %a1
ret <4 x i64> %2
}
define <16 x i8> @stack_fold_psubsb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubsb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubsb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.ssub.sat.v16i8(<16 x i8> %a0, <16 x i8> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.ssub.sat.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
define <32 x i8> @stack_fold_psubsb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubsb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubsb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.ssub.sat.v32i8(<32 x i8> %a0, <32 x i8> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.ssub.sat.v32i8(<32 x i8>, <32 x i8>) nounwind readnone
define <8 x i16> @stack_fold_psubsw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubsw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubsw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psubsw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubsw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubsw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.ssub.sat.v16i16(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.ssub.sat.v16i16(<16 x i16>, <16 x i16>) nounwind readnone
define <16 x i8> @stack_fold_psubusb(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubusb:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubusb {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i8> @llvm.usub.sat.v16i8(<16 x i8> %a0, <16 x i8> %a1)
ret <16 x i8> %2
}
declare <16 x i8> @llvm.usub.sat.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
define <32 x i8> @stack_fold_psubusb_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_psubusb_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubusb {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <32 x i8> @llvm.usub.sat.v32i8(<32 x i8> %a0, <32 x i8> %a1)
ret <32 x i8> %2
}
declare <32 x i8> @llvm.usub.sat.v32i8(<32 x i8>, <32 x i8>) nounwind readnone
define <8 x i16> @stack_fold_psubusw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubusw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubusw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> %a0, <8 x i16> %a1)
ret <8 x i16> %2
}
declare <8 x i16> @llvm.usub.sat.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
define <16 x i16> @stack_fold_psubusw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubusw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubusw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = call <16 x i16> @llvm.usub.sat.v16i16(<16 x i16> %a0, <16 x i16> %a1)
ret <16 x i16> %2
}
declare <16 x i16> @llvm.usub.sat.v16i16(<16 x i16>, <16 x i16>) nounwind readnone
define <8 x i16> @stack_fold_psubw(<8 x i16> %a0, <8 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <8 x i16> %a0, %a1
ret <8 x i16> %2
}
define <16 x i16> @stack_fold_psubw_ymm(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: stack_fold_psubw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpsubw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = sub <16 x i16> %a0, %a1
ret <16 x i16> %2
}
define <16 x i8> @stack_fold_punpckhbw(<16 x i8> %a0, <16 x i8> %a1) {
; CHECK-LABEL: stack_fold_punpckhbw:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 = xmm0[8],mem[8],xmm0[9],mem[9],xmm0[10],mem[10],xmm0[11],mem[11],xmm0[12],mem[12],xmm0[13],mem[13],xmm0[14],mem[14],xmm0[15],mem[15]
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> %a1, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
ret <16 x i8> %2
}
define <16 x i8> @stack_fold_punpckhbw_mask(ptr %passthru, <16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_punpckhbw_mask:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %xmm2
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm2 {%k1} # 16-byte Folded Reload
; CHECK-NEXT: # xmm2 {%k1} = xmm0[8],mem[8],xmm0[9],mem[9],xmm0[10],mem[10],xmm0[11],mem[11],xmm0[12],mem[12],xmm0[13],mem[13],xmm0[14],mem[14],xmm0[15],mem[15]
; CHECK-NEXT: vmovdqa %xmm2, %xmm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> %a1, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
%3 = bitcast i16 %mask to <16 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <16 x i8>, ptr %passthru
%5 = select <16 x i1> %3, <16 x i8> %2, <16 x i8> %4
ret <16 x i8> %5
}
define <16 x i8> @stack_fold_punpckhbw_maskz(<16 x i8> %passthru, <16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
; CHECK-LABEL: stack_fold_punpckhbw_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm0 {%k1} {z} # 16-byte Folded Reload
; CHECK-NEXT: # xmm0 {%k1} {z} = xmm1[8],mem[8],xmm1[9],mem[9],xmm1[10],mem[10],xmm1[11],mem[11],xmm1[12],mem[12],xmm1[13],mem[13],xmm1[14],mem[14],xmm1[15],mem[15]
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <16 x i8> %a0, <16 x i8> %a1, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
%3 = bitcast i16 %mask to <16 x i1>
%4 = select <16 x i1> %3, <16 x i8> %2, <16 x i8> zeroinitializer
ret <16 x i8> %4
}
define <32 x i8> @stack_fold_punpckhbw_ymm(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: stack_fold_punpckhbw_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 = ymm0[8],mem[8],ymm0[9],mem[9],ymm0[10],mem[10],ymm0[11],mem[11],ymm0[12],mem[12],ymm0[13],mem[13],ymm0[14],mem[14],ymm0[15],mem[15],ymm0[24],mem[24],ymm0[25],mem[25],ymm0[26],mem[26],ymm0[27],mem[27],ymm0[28],mem[28],ymm0[29],mem[29],ymm0[30],mem[30],ymm0[31],mem[31]
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a0, <32 x i8> %a1, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
ret <32 x i8> %2
}
define <32 x i8> @stack_fold_punpckhbw_mask_ymm(ptr %passthru, <32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_punpckhbw_mask_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %esi, %k1
; CHECK-NEXT: vmovdqa (%rdi), %ymm2
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm2 {%k1} # 32-byte Folded Reload
; CHECK-NEXT: # ymm2 {%k1} = ymm0[8],mem[8],ymm0[9],mem[9],ymm0[10],mem[10],ymm0[11],mem[11],ymm0[12],mem[12],ymm0[13],mem[13],ymm0[14],mem[14],ymm0[15],mem[15],ymm0[24],mem[24],ymm0[25],mem[25],ymm0[26],mem[26],ymm0[27],mem[27],ymm0[28],mem[28],ymm0[29],mem[29],ymm0[30],mem[30],ymm0[31],mem[31]
; CHECK-NEXT: vmovdqa %ymm2, %ymm0
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a0, <32 x i8> %a1, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
%3 = bitcast i32 %mask to <32 x i1>
; load needed to keep the operation from being scheduled about the asm block
%4 = load <32 x i8>, ptr %passthru
%5 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> %4
ret <32 x i8> %5
}
define <32 x i8> @stack_fold_punpckhbw_maskz_ymm(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
; CHECK-LABEL: stack_fold_punpckhbw_maskz_ymm:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vpunpckhbw {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = ymm0[8],mem[8],ymm0[9],mem[9],ymm0[10],mem[10],ymm0[11],mem[11],ymm0[12],mem[12],ymm0[13],mem[13],ymm0[14],mem[14],ymm0[15],mem[15],ymm0[24],mem[24],ymm0[25],mem[25],ymm0[26],mem[26],ymm0[27],mem[27],ymm0[28],mem[28],ymm0[29],mem[29],ymm0[30],mem[30],ymm0[31],mem[31]
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <32 x i8> %a0, <32 x i8> %a1, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
%3 = bitcast i32 %mask to <32 x i1>
%4 = select <32 x i1> %3, <32 x i8> %2, <32 x i8> zeroinitializer
ret <32 x i8> %4
}
define <4 x i64> @stack_fold_shufi64x2_maskz(<4 x i64> %a, <4 x i64> %b, i8 %mask) {
; CHECK-LABEL: stack_fold_shufi64x2_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vshufi64x2 $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = ymm0[2,3],mem[0,1]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
%3 = bitcast i8 %mask to <8 x i1>
%4 = shufflevector <8 x i1> %3, <8 x i1> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%5 = select <4 x i1> %4, <4 x i64> %2, <4 x i64> zeroinitializer
ret <4 x i64> %5
}
define <8 x i32> @stack_fold_shufi32x4_maskz(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: stack_fold_shufi32x4_maskz:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: kmovd %edi, %k1
; CHECK-NEXT: vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; CHECK-NEXT: vshufi32x4 $1, {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %ymm0 {%k1} {z} # 32-byte Folded Reload
; CHECK-NEXT: # ymm0 {%k1} {z} = ymm0[4,5,6,7],mem[0,1,2,3]
; CHECK-NEXT: retq
%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},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
%3 = bitcast i8 %mask to <8 x i1>
%4 = select <8 x i1> %3, <8 x i32> %2, <8 x i32> zeroinitializer
ret <8 x i32> %4
}
declare <4 x i32> @llvm.x86.avx512.conflict.d.128(<4 x i32>)
declare <8 x i32> @llvm.x86.avx512.conflict.d.256(<8 x i32>)
declare <2 x i64> @llvm.x86.avx512.conflict.q.128(<2 x i64>)
declare <4 x i64> @llvm.x86.avx512.conflict.q.256(<4 x i64>)
declare <4 x i32> @llvm.ctlz.v4i32(<4 x i32>, i1)
declare <8 x i32> @llvm.ctlz.v8i32(<8 x i32>, i1)
declare <2 x i64> @llvm.ctlz.v2i64(<2 x i64>, i1)
declare <4 x i64> @llvm.ctlz.v4i64(<4 x i64>, i1)