[AVX-512] Add masked shift instructions to load folding tables.

This adds the masked versions of everything, but the shift by immediate instructions.

llvm-svn: 294286

1 files changed, 42 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/X86/stack-folding-int-avx512.ll b/llvm/test/CodeGen/X86/stack-folding-int-avx512.ll
index ba384a2bff8..3201f7bda3a 100644
--- a/llvm/test/CodeGen/X86/stack-folding-int-avx512.ll
+++ b/llvm/test/CodeGen/X86/stack-folding-int-avx512.ll
@@ -1047,6 +1047,27 @@ define <16 x i32> @stack_fold_pslld(<16 x i32> %a0, <4 x i32> %a1) {
 }
 declare <16 x i32> @llvm.x86.avx512.psll.d.512(<16 x i32>, <4 x i32>) nounwind readnone
 
+define <16 x i32> @stack_fold_pslld_mask(<16 x i32>* %passthru, <16 x i32> %a0, <4 x i32> %a1, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_pslld_mask
+  ;CHECK:       vpslld {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x i32> @llvm.x86.avx512.psll.d.512(<16 x i32> %a0, <4 x i32> %a1)
+  %3 = bitcast i16 %mask to <16 x i1>
+  %4 = load <16 x i32>, <16 x i32>* %passthru
+  %5 = select <16 x i1> %3, <16 x i32> %2, <16 x i32> %4
+  ret <16 x i32> %5
+}
+
+define <16 x i32> @stack_fold_pslld_maskz(<16 x i32> %a0, <4 x i32> %a1, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_pslld_maskz
+  ;CHECK:       vpslld {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x i32> @llvm.x86.avx512.psll.d.512(<16 x i32> %a0, <4 x i32> %a1)
+  %3 = bitcast i16 %mask to <16 x i1>
+  %4 = select <16 x i1> %3, <16 x i32> %2, <16 x i32> zeroinitializer
+  ret <16 x i32> %4
+}
+
 define <64 x i8> @stack_fold_pslldq(<64 x i8> %a, <64 x i8> %b) {
   ;CHECK-LABEL: stack_fold_pslldq
   ;CHECK:       vpslldq $1, {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload
@@ -1073,6 +1094,27 @@ define <16 x i32> @stack_fold_psllvd(<16 x i32> %a0, <16 x i32> %a1) {
 }
 declare <16 x i32> @llvm.x86.avx512.psllv.d.512(<16 x i32>, <16 x i32>) nounwind readnone
 
+define <16 x i32> @stack_fold_psllvd_mask(<16 x i32>* %passthru, <16 x i32> %a0, <16 x i32> %a1, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_psllvd_mask
+  ;CHECK:       vpsllvd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x i32> @llvm.x86.avx512.psllv.d.512(<16 x i32> %a0, <16 x i32> %a1)
+  %3 = bitcast i16 %mask to <16 x i1>
+  %4 = load <16 x i32>, <16 x i32>* %passthru
+  %5 = select <16 x i1> %3, <16 x i32> %2, <16 x i32> %4
+  ret <16 x i32> %5
+}
+
+define <16 x i32> @stack_fold_psllvd_maskz(<16 x i32> %a0, <16 x i32> %a1, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_psllvd_maskz
+  ;CHECK:       vpsllvd {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x i32> @llvm.x86.avx512.psllv.d.512(<16 x i32> %a0, <16 x i32> %a1)
+  %3 = bitcast i16 %mask to <16 x i1>
+  %4 = select <16 x i1> %3, <16 x i32> %2, <16 x i32> zeroinitializer
+  ret <16 x i32> %4
+}
+
 define <8 x i64> @stack_fold_psllvq(<8 x i64> %a0, <8 x i64> %a1) {
   ;CHECK-LABEL: stack_fold_psllvq
   ;CHECK:       vpsllvq {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload