summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll
diff options
context:
space:
mode:
Diffstat (limited to 'llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll')
-rw-r--r--llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll80
1 files changed, 30 insertions, 50 deletions
diff --git a/llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll b/llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll
index 51d056f2049..d08597b5195 100644
--- a/llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll
+++ b/llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll
@@ -1,36 +1,26 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+sse4a | FileCheck %s --check-prefix=X32
-; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+sse4a,+avx | FileCheck %s --check-prefix=X32
-; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4a | FileCheck %s --check-prefix=X64
-; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4a,+avx | FileCheck %s --check-prefix=X64
+; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+sse4a | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+sse4a,+avx | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4a | FileCheck %s --check-prefixes=CHECK,X64
+; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+sse4a,+avx | FileCheck %s --check-prefixes=CHECK,X64
; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/sse4a-builtins.c
define <2 x i64> @test_mm_extracti_si64(<2 x i64> %x) {
-; X32-LABEL: test_mm_extracti_si64:
-; X32: # %bb.0:
-; X32-NEXT: extrq $2, $3, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm_extracti_si64:
-; X64: # %bb.0:
-; X64-NEXT: extrq $2, $3, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm_extracti_si64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: extrq $2, $3, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = call <2 x i64> @llvm.x86.sse4a.extrqi(<2 x i64> %x, i8 3, i8 2)
ret <2 x i64> %res
}
declare <2 x i64> @llvm.x86.sse4a.extrqi(<2 x i64>, i8, i8) nounwind readnone
define <2 x i64> @test_mm_extract_si64(<2 x i64> %x, <2 x i64> %y) {
-; X32-LABEL: test_mm_extract_si64:
-; X32: # %bb.0:
-; X32-NEXT: extrq %xmm1, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm_extract_si64:
-; X64: # %bb.0:
-; X64-NEXT: extrq %xmm1, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm_extract_si64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: extrq %xmm1, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%bc = bitcast <2 x i64> %y to <16 x i8>
%res = call <2 x i64> @llvm.x86.sse4a.extrq(<2 x i64> %x, <16 x i8> %bc)
ret <2 x i64> %res
@@ -38,41 +28,31 @@ define <2 x i64> @test_mm_extract_si64(<2 x i64> %x, <2 x i64> %y) {
declare <2 x i64> @llvm.x86.sse4a.extrq(<2 x i64>, <16 x i8>) nounwind readnone
define <2 x i64> @test_mm_inserti_si64(<2 x i64> %x, <2 x i64> %y) {
-; X32-LABEL: test_mm_inserti_si64:
-; X32: # %bb.0:
-; X32-NEXT: insertq $6, $5, %xmm1, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm_inserti_si64:
-; X64: # %bb.0:
-; X64-NEXT: insertq $6, $5, %xmm1, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm_inserti_si64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: insertq $6, $5, %xmm1, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %x, <2 x i64> %y, i8 5, i8 6)
ret <2 x i64> %res
}
declare <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64>, <2 x i64>, i8, i8) nounwind readnone
define <2 x i64> @test_mm_insert_si64(<2 x i64> %x, <2 x i64> %y) {
-; X32-LABEL: test_mm_insert_si64:
-; X32: # %bb.0:
-; X32-NEXT: insertq %xmm1, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm_insert_si64:
-; X64: # %bb.0:
-; X64-NEXT: insertq %xmm1, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm_insert_si64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: insertq %xmm1, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = call <2 x i64> @llvm.x86.sse4a.insertq(<2 x i64> %x, <2 x i64> %y)
ret <2 x i64> %res
}
declare <2 x i64> @llvm.x86.sse4a.insertq(<2 x i64>, <2 x i64>) nounwind readnone
define void @test_stream_sd(double* %p, <2 x double> %a) {
-; X32-LABEL: test_stream_sd:
-; X32: # %bb.0:
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: movntsd %xmm0, (%eax)
-; X32-NEXT: retl
+; X86-LABEL: test_stream_sd:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: movntsd %xmm0, (%eax)
+; X86-NEXT: retl
;
; X64-LABEL: test_stream_sd:
; X64: # %bb.0:
@@ -84,11 +64,11 @@ define void @test_stream_sd(double* %p, <2 x double> %a) {
}
define void @test_mm_stream_ss(float* %p, <4 x float> %a) {
-; X32-LABEL: test_mm_stream_ss:
-; X32: # %bb.0:
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: movntss %xmm0, (%eax)
-; X32-NEXT: retl
+; X86-LABEL: test_mm_stream_ss:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: movntss %xmm0, (%eax)
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_stream_ss:
; X64: # %bb.0:
OpenPOWER on IntegriCloud