[X86] Add test coverage for #162812

[RKSimon]

No description provided.

[llvmbot]

@llvm/pr-subscribers-backend-x86

Author: Simon Pilgrim (RKSimon)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/162878.diff

1 Files Affected:

• (added) llvm/test/CodeGen/X86/pr162812.ll (+98)

diff --git a/llvm/test/CodeGen/X86/pr162812.ll b/llvm/test/CodeGen/X86/pr162812.ll
new file mode 100644
index 0000000000000..4ea3101ce8456
--- /dev/null
+++ b/llvm/test/CodeGen/X86/pr162812.ll
@@ -0,0 +1,98 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
+; RUN: llc < %s -mtriple=x86_64-- -mcpu=x86-64    | FileCheck %s --check-prefixes=SSE2
+; RUN: llc < %s -mtriple=x86_64-- -mcpu=x86-64-v2 | FileCheck %s --check-prefixes=SSE42
+; RUN: llc < %s -mtriple=x86_64-- -mcpu=x86-64-v3 | FileCheck %s --check-prefixes=AVX2
+; RUN: llc < %s -mtriple=x86_64-- -mcpu=x86-64-v4 | FileCheck %s --check-prefixes=AVX512
+
+define <32 x i8> @PR162812(<32 x i8> %a, <32 x i8> %mask) {
+; SSE2-LABEL: PR162812:
+; SSE2:       # %bb.0:
+; SSE2-NEXT:    psrlw $2, %xmm2
+; SSE2-NEXT:    movdqa {{.*#+}} xmm4 = [8224,8224,8224,8224,8224,8224,8224,8224]
+; SSE2-NEXT:    pand %xmm4, %xmm2
+; SSE2-NEXT:    paddb %xmm2, %xmm2
+; SSE2-NEXT:    paddb %xmm2, %xmm2
+; SSE2-NEXT:    pxor %xmm5, %xmm5
+; SSE2-NEXT:    pxor %xmm6, %xmm6
+; SSE2-NEXT:    pcmpgtb %xmm2, %xmm6
+; SSE2-NEXT:    movdqa %xmm6, %xmm2
+; SSE2-NEXT:    pandn %xmm0, %xmm2
+; SSE2-NEXT:    paddb %xmm0, %xmm0
+; SSE2-NEXT:    pand %xmm6, %xmm0
+; SSE2-NEXT:    por %xmm2, %xmm0
+; SSE2-NEXT:    psrlw $2, %xmm3
+; SSE2-NEXT:    pand %xmm4, %xmm3
+; SSE2-NEXT:    paddb %xmm3, %xmm3
+; SSE2-NEXT:    paddb %xmm3, %xmm3
+; SSE2-NEXT:    pcmpgtb %xmm3, %xmm5
+; SSE2-NEXT:    movdqa %xmm5, %xmm2
+; SSE2-NEXT:    pandn %xmm1, %xmm2
+; SSE2-NEXT:    paddb %xmm1, %xmm1
+; SSE2-NEXT:    pand %xmm5, %xmm1
+; SSE2-NEXT:    por %xmm2, %xmm1
+; SSE2-NEXT:    retq
+;
+; SSE42-LABEL: PR162812:
+; SSE42:       # %bb.0:
+; SSE42-NEXT:    movdqa %xmm2, %xmm5
+; SSE42-NEXT:    movdqa %xmm0, %xmm2
+; SSE42-NEXT:    movdqa %xmm0, %xmm6
+; SSE42-NEXT:    psllw $2, %xmm6
+; SSE42-NEXT:    movdqa {{.*#+}} xmm7 = [252,252,252,252,252,252,252,252,252,252,252,252,252,252,252,252]
+; SSE42-NEXT:    pand %xmm7, %xmm6
+; SSE42-NEXT:    psrlw $2, %xmm5
+; SSE42-NEXT:    movdqa {{.*#+}} xmm4 = [8224,8224,8224,8224,8224,8224,8224,8224]
+; SSE42-NEXT:    pand %xmm4, %xmm5
+; SSE42-NEXT:    paddb %xmm5, %xmm5
+; SSE42-NEXT:    movdqa %xmm5, %xmm0
+; SSE42-NEXT:    pblendvb %xmm0, %xmm6, %xmm2
+; SSE42-NEXT:    movdqa %xmm2, %xmm6
+; SSE42-NEXT:    paddb %xmm2, %xmm6
+; SSE42-NEXT:    paddb %xmm5, %xmm5
+; SSE42-NEXT:    movdqa %xmm5, %xmm0
+; SSE42-NEXT:    pblendvb %xmm0, %xmm6, %xmm2
+; SSE42-NEXT:    movdqa %xmm1, %xmm5
+; SSE42-NEXT:    psllw $2, %xmm5
+; SSE42-NEXT:    pand %xmm7, %xmm5
+; SSE42-NEXT:    psrlw $2, %xmm3
+; SSE42-NEXT:    pand %xmm3, %xmm4
+; SSE42-NEXT:    paddb %xmm4, %xmm4
+; SSE42-NEXT:    movdqa %xmm4, %xmm0
+; SSE42-NEXT:    pblendvb %xmm0, %xmm5, %xmm1
+; SSE42-NEXT:    movdqa %xmm1, %xmm3
+; SSE42-NEXT:    paddb %xmm1, %xmm3
+; SSE42-NEXT:    paddb %xmm4, %xmm4
+; SSE42-NEXT:    movdqa %xmm4, %xmm0
+; SSE42-NEXT:    pblendvb %xmm0, %xmm3, %xmm1
+; SSE42-NEXT:    movdqa %xmm2, %xmm0
+; SSE42-NEXT:    retq
+;
+; AVX2-LABEL: PR162812:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpsllw $2, %ymm0, %ymm2
+; AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm2
+; AVX2-NEXT:    vpsrlw $2, %ymm1, %ymm1
+; AVX2-NEXT:    vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm1
+; AVX2-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX2-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpaddb %ymm0, %ymm0, %ymm2
+; AVX2-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX2-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: PR162812:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpsllw $2, %ymm0, %ymm2
+; AVX512-NEXT:    vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm2, %ymm2
+; AVX512-NEXT:    vpsrlw $2, %ymm1, %ymm1
+; AVX512-NEXT:    vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm1, %ymm1
+; AVX512-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512-NEXT:    vpaddb %ymm0, %ymm0, %ymm2
+; AVX512-NEXT:    vpaddb %ymm1, %ymm1, %ymm1
+; AVX512-NEXT:    vpblendvb %ymm1, %ymm2, %ymm0, %ymm0
+; AVX512-NEXT:    retq
+  %1 = lshr <32 x i8> %mask, splat (i8 7)
+  %ret = shl <32 x i8> %a, %1
+  ret <32 x i8> %ret
+}