[X86] Add 512-bit test cases to machine-combiner-int-vec.ll. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@362549 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/test/CodeGen/X86/machine-combiner-int-vec.ll b/test/CodeGen/X86/machine-combiner-int-vec.ll
index 5aef93a84f71fa0eb2944f5112469a51f8d3439d..81844f6b5918a9b7b877652f2cfa6df3f51bf0bd 100644 (file)
--- a/test/CodeGen/X86/machine-combiner-int-vec.ll
+++ b/test/CodeGen/X86/machine-combiner-int-vec.ll
@@ -1,5 +1,7 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
 ; RUN: llc -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=sse2 -machine-combiner-verify-pattern-order=true < %s | FileCheck %s --check-prefix=SSE
-; RUN: llc -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=avx2 -machine-combiner-verify-pattern-order=true < %s | FileCheck %s --check-prefix=AVX
+; RUN: llc -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=avx2 -machine-combiner-verify-pattern-order=true < %s | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
+; RUN: llc -mtriple=x86_64-unknown-unknown -mcpu=x86-64 -mattr=avx512vl -machine-combiner-verify-pattern-order=true < %s | FileCheck %s --check-prefix=AVX --check-prefix=AVX512
 
 ; Verify that 128-bit vector logical ops are reassociated.
 
@@ -69,6 +71,16 @@ define <4 x i32> @reassociate_xor_v4i32(<4 x i32> %x0, <4 x i32> %x1, <4 x i32>
 ; Verify that 256-bit vector logical ops are reassociated.
 
 define <8 x i32> @reassociate_and_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, <8 x i32> %x3) {
+; SSE-LABEL: reassociate_and_v8i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm3, %xmm1
+; SSE-NEXT:    paddd %xmm2, %xmm0
+; SSE-NEXT:    pand %xmm6, %xmm4
+; SSE-NEXT:    pand %xmm4, %xmm0
+; SSE-NEXT:    pand %xmm7, %xmm5
+; SSE-NEXT:    pand %xmm5, %xmm1
+; SSE-NEXT:    retq
+;
 ; AVX-LABEL: reassociate_and_v8i32:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpaddd %ymm1, %ymm0, %ymm0
@@ -83,6 +95,16 @@ define <8 x i32> @reassociate_and_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>
 }
 
 define <8 x i32> @reassociate_or_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, <8 x i32> %x3) {
+; SSE-LABEL: reassociate_or_v8i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm3, %xmm1
+; SSE-NEXT:    paddd %xmm2, %xmm0
+; SSE-NEXT:    por %xmm6, %xmm4
+; SSE-NEXT:    por %xmm4, %xmm0
+; SSE-NEXT:    por %xmm7, %xmm5
+; SSE-NEXT:    por %xmm5, %xmm1
+; SSE-NEXT:    retq
+;
 ; AVX-LABEL: reassociate_or_v8i32:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpaddd %ymm1, %ymm0, %ymm0
@@ -97,6 +119,16 @@ define <8 x i32> @reassociate_or_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %
 }
 
 define <8 x i32> @reassociate_xor_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, <8 x i32> %x3) {
+; SSE-LABEL: reassociate_xor_v8i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm3, %xmm1
+; SSE-NEXT:    paddd %xmm2, %xmm0
+; SSE-NEXT:    pxor %xmm6, %xmm4
+; SSE-NEXT:    pxor %xmm4, %xmm0
+; SSE-NEXT:    pxor %xmm7, %xmm5
+; SSE-NEXT:    pxor %xmm5, %xmm1
+; SSE-NEXT:    retq
+;
 ; AVX-LABEL: reassociate_xor_v8i32:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpaddd %ymm1, %ymm0, %ymm0
@@ -110,3 +142,125 @@ define <8 x i32> @reassociate_xor_v8i32(<8 x i32> %x0, <8 x i32> %x1, <8 x i32>
   ret <8 x i32> %t2
 }
 
+
+; Verify that 512-bit vector logical ops are reassociated.
+
+define <16 x i32> @reassociate_and_v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, <16 x i32> %x3) {
+; SSE-LABEL: reassociate_and_v16i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm4, %xmm0
+; SSE-NEXT:    paddd %xmm5, %xmm1
+; SSE-NEXT:    paddd %xmm6, %xmm2
+; SSE-NEXT:    paddd %xmm7, %xmm3
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    pand {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    retq
+;
+; AVX2-LABEL: reassociate_and_v16i32:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpaddd %ymm3, %ymm1, %ymm1
+; AVX2-NEXT:    vpaddd %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpand %ymm6, %ymm4, %ymm2
+; AVX2-NEXT:    vpand %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpand %ymm7, %ymm5, %ymm2
+; AVX2-NEXT:    vpand %ymm2, %ymm1, %ymm1
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: reassociate_and_v16i32:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    vpandd %zmm3, %zmm2, %zmm1
+; AVX512-NEXT:    vpandd %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    retq
+
+  %t0 = add <16 x i32> %x0, %x1
+  %t1 = and <16 x i32> %x2, %t0
+  %t2 = and <16 x i32> %x3, %t1
+  ret <16 x i32> %t2
+}
+
+define <16 x i32> @reassociate_or_v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, <16 x i32> %x3) {
+; SSE-LABEL: reassociate_or_v16i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm4, %xmm0
+; SSE-NEXT:    paddd %xmm5, %xmm1
+; SSE-NEXT:    paddd %xmm6, %xmm2
+; SSE-NEXT:    paddd %xmm7, %xmm3
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    por {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    retq
+;
+; AVX2-LABEL: reassociate_or_v16i32:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpaddd %ymm3, %ymm1, %ymm1
+; AVX2-NEXT:    vpaddd %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpor %ymm6, %ymm4, %ymm2
+; AVX2-NEXT:    vpor %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpor %ymm7, %ymm5, %ymm2
+; AVX2-NEXT:    vpor %ymm2, %ymm1, %ymm1
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: reassociate_or_v16i32:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    vpord %zmm3, %zmm2, %zmm1
+; AVX512-NEXT:    vpord %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    retq
+
+  %t0 = add <16 x i32> %x0, %x1
+  %t1 = or <16 x i32> %x2, %t0
+  %t2 = or <16 x i32> %x3, %t1
+  ret <16 x i32> %t2
+}
+
+define <16 x i32> @reassociate_xor_v16i32(<16 x i32> %x0, <16 x i32> %x1, <16 x i32> %x2, <16 x i32> %x3) {
+; SSE-LABEL: reassociate_xor_v16i32:
+; SSE:       # %bb.0:
+; SSE-NEXT:    paddd %xmm4, %xmm0
+; SSE-NEXT:    paddd %xmm5, %xmm1
+; SSE-NEXT:    paddd %xmm6, %xmm2
+; SSE-NEXT:    paddd %xmm7, %xmm3
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm0
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm1
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm2
+; SSE-NEXT:    pxor {{[0-9]+}}(%rsp), %xmm3
+; SSE-NEXT:    retq
+;
+; AVX2-LABEL: reassociate_xor_v16i32:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    vpaddd %ymm3, %ymm1, %ymm1
+; AVX2-NEXT:    vpaddd %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpxor %ymm6, %ymm4, %ymm2
+; AVX2-NEXT:    vpxor %ymm2, %ymm0, %ymm0
+; AVX2-NEXT:    vpxor %ymm7, %ymm5, %ymm2
+; AVX2-NEXT:    vpxor %ymm2, %ymm1, %ymm1
+; AVX2-NEXT:    retq
+;
+; AVX512-LABEL: reassociate_xor_v16i32:
+; AVX512:       # %bb.0:
+; AVX512-NEXT:    vpaddd %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    vpxord %zmm3, %zmm2, %zmm1
+; AVX512-NEXT:    vpxord %zmm1, %zmm0, %zmm0
+; AVX512-NEXT:    retq
+
+  %t0 = add <16 x i32> %x0, %x1
+  %t1 = xor <16 x i32> %x2, %t0
+  %t2 = xor <16 x i32> %x3, %t1
+  ret <16 x i32> %t2
+}