[X86] Add AVX2 test coverage to TESTPS/TESTPD combines

diff --git a/llvm/test/CodeGen/X86/combine-testpd.ll b/llvm/test/CodeGen/X86/combine-testpd.ll
index 53e8193..0e0df82 100644 (file)
--- a/llvm/test/CodeGen/X86/combine-testpd.ll
+++ b/llvm/test/CodeGen/X86/combine-testpd.ll
@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx  | FileCheck %s --check-prefixes=CHECK,AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX2
 
 ;
 ; testz(~X,Y) -> testc(X,Y)
@@ -188,18 +189,26 @@ define i32 @testpdz_256_signbit(<4 x double> %c, <4 x double> %d, i32 %a, i32 %b
 }
 
 define i32 @testpdnzc_256_signbit_multiuse(<4 x double> %c, i32 %a, i32 %b) {
-; CHECK-LABEL: testpdnzc_256_signbit_multiuse:
-; CHECK:       # %bb.0:
-; CHECK-NEXT:    movl %edi, %eax
-; CHECK-NEXT:    vextractf128 $1, %ymm0, %xmm1
-; CHECK-NEXT:    vpxor %xmm2, %xmm2, %xmm2
-; CHECK-NEXT:    vpcmpgtq %xmm1, %xmm2, %xmm1
-; CHECK-NEXT:    vpcmpgtq %xmm0, %xmm2, %xmm0
-; CHECK-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
-; CHECK-NEXT:    vtestpd %ymm0, %ymm0
-; CHECK-NEXT:    cmovnel %esi, %eax
-; CHECK-NEXT:    vzeroupper
-; CHECK-NEXT:    retq
+; AVX1-LABEL: testpdnzc_256_signbit_multiuse:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    movl %edi, %eax
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm1
+; AVX1-NEXT:    vpxor %xmm2, %xmm2, %xmm2
+; AVX1-NEXT:    vpcmpgtq %xmm1, %xmm2, %xmm1
+; AVX1-NEXT:    vpcmpgtq %xmm0, %xmm2, %xmm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT:    vtestpd %ymm0, %ymm0
+; AVX1-NEXT:    cmovnel %esi, %eax
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: testpdnzc_256_signbit_multiuse:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    movl %edi, %eax
+; AVX2-NEXT:    vtestpd %ymm0, %ymm0
+; AVX2-NEXT:    cmovnel %esi, %eax
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
   %t0 = bitcast <4 x double> %c to <4 x i64>
   %t1 = icmp sgt <4 x i64> zeroinitializer, %t0
   %t2 = sext <4 x i1> %t1 to <4 x i64>

diff --git a/llvm/test/CodeGen/X86/combine-testps.ll b/llvm/test/CodeGen/X86/combine-testps.ll
index 6158fc5..2817ffb 100644 (file)
--- a/llvm/test/CodeGen/X86/combine-testps.ll
+++ b/llvm/test/CodeGen/X86/combine-testps.ll
@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx  | FileCheck %s --check-prefixes=CHECK,AVX1
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX2
 
 ;
 ; testz(~X,Y) -> testc(X,Y)
@@ -188,17 +189,25 @@ define i32 @testpsnzc_256_signbit(<8 x float> %c, <8 x float> %d, i32 %a, i32 %b
 }
 
 define i32 @testpsc_256_signbit_multiuse(<8 x float> %c, i32 %a, i32 %b) {
-; CHECK-LABEL: testpsc_256_signbit_multiuse:
-; CHECK:       # %bb.0:
-; CHECK-NEXT:    movl %edi, %eax
-; CHECK-NEXT:    vpsrad $31, %xmm0, %xmm1
-; CHECK-NEXT:    vextractf128 $1, %ymm0, %xmm0
-; CHECK-NEXT:    vpsrad $31, %xmm0, %xmm0
-; CHECK-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
-; CHECK-NEXT:    vtestps %ymm0, %ymm0
-; CHECK-NEXT:    cmovnel %esi, %eax
-; CHECK-NEXT:    vzeroupper
-; CHECK-NEXT:    retq
+; AVX1-LABEL: testpsc_256_signbit_multiuse:
+; AVX1:       # %bb.0:
+; AVX1-NEXT:    movl %edi, %eax
+; AVX1-NEXT:    vpsrad $31, %xmm0, %xmm1
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT:    vpsrad $31, %xmm0, %xmm0
+; AVX1-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
+; AVX1-NEXT:    vtestps %ymm0, %ymm0
+; AVX1-NEXT:    cmovnel %esi, %eax
+; AVX1-NEXT:    vzeroupper
+; AVX1-NEXT:    retq
+;
+; AVX2-LABEL: testpsc_256_signbit_multiuse:
+; AVX2:       # %bb.0:
+; AVX2-NEXT:    movl %edi, %eax
+; AVX2-NEXT:    vtestps %ymm0, %ymm0
+; AVX2-NEXT:    cmovnel %esi, %eax
+; AVX2-NEXT:    vzeroupper
+; AVX2-NEXT:    retq
   %t0 = bitcast <8 x float> %c to <8 x i32>
   %t1 = ashr <8 x i32> %t0, <i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31>
   %t2 = bitcast <8 x i32> %t1 to <8 x float>