CodeGen peephole: fold redundant phys reg copies

[oota-llvm.git] / test / CodeGen / X86 / fma.ll

diff --git a/test/CodeGen/X86/fma.ll b/test/CodeGen/X86/fma.ll
index 30bdf9535f98bda1b8ef2970c280703c436d9482..b91479cda871508dc09df11284cad7b97afa95bb 100644 (file)
--- a/test/CodeGen/X86/fma.ll
+++ b/test/CodeGen/X86/fma.ll
@@ -1,39 +1,47 @@
-; RUN: llc < %s -mtriple=i386-apple-darwin10  -mattr=+fma,-fma4  | FileCheck %s --check-prefix=CHECK-FMA-INST
-; RUN: llc < %s -mtriple=i386-apple-darwin10  -mattr=-fma,-fma4  | FileCheck %s --check-prefix=CHECK-FMA-CALL
-; RUN: llc < %s -mtriple=x86_64-apple-darwin10 -mattr=+fma,-fma4 | FileCheck %s --check-prefix=CHECK-FMA-INST
-; RUN: llc < %s -mtriple=x86_64-apple-darwin10  -mattr=-fma,-fma4 | FileCheck %s --check-prefix=CHECK-FMA-CALL
-; RUN: llc < %s -mcpu=bdver2 -mattr=-fma4  | FileCheck %s --check-prefix=CHECK-FMA-INST
-; RUN: llc < %s -mcpu=bdver2 -mattr=-fma,-fma4 | FileCheck %s --check-prefix=CHECK-FMA-CALL
+; RUN: llc < %s -mtriple=i386-apple-darwin10  -mattr=+fma,-fma4  | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-INST
+; RUN: llc < %s -mtriple=i386-apple-darwin10  -mattr=-fma,-fma4  | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-CALL
+; RUN: llc < %s -mtriple=x86_64-apple-darwin10 -mattr=+fma,-fma4 | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-INST
+; RUN: llc < %s -mtriple=x86_64-apple-darwin10  -mattr=-fma,-fma4 | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-CALL
+; RUN: llc < %s -mtriple=x86_64-apple-darwin10  -mattr=+avx512f,-fma,-fma4 | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-INST
+; RUN: llc < %s -march=x86 -mcpu=bdver2 -mattr=-fma4  | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-INST
+; RUN: llc < %s -march=x86 -mcpu=bdver2 -mattr=-fma,-fma4 | FileCheck %s --check-prefix=CHECK --check-prefix=CHECK-FMA-CALL
 
-; CHECK: test_f32
+; CHECK-LABEL: test_f32:
 ; CHECK-FMA-INST: vfmadd213ss
 ; CHECK-FMA-CALL: fmaf
-
-define float @test_f32(float %a, float %b, float %c) nounwind readnone ssp {
+define float @test_f32(float %a, float %b, float %c) #0 {
 entry:
-  %call = tail call float @llvm.fma.f32(float %a, float %b, float %c) nounwind readnone
+  %call = call float @llvm.fma.f32(float %a, float %b, float %c)
   ret float %call
 }
 
-; CHECK: test_f64
+; CHECK-LABEL: test_f64:
 ; CHECK-FMA-INST: vfmadd213sd
 ; CHECK-FMA-CALL: fma
-
-define double @test_f64(double %a, double %b, double %c) nounwind readnone ssp {
+define double @test_f64(double %a, double %b, double %c) #0 {
 entry:
-  %call = tail call double @llvm.fma.f64(double %a, double %b, double %c) nounwind readnone
+  %call = call double @llvm.fma.f64(double %a, double %b, double %c)
   ret double %call
 }
 
-; CHECK: test_f80
+; CHECK-LABEL: test_f80:
 ; CHECK: fmal
-
-define x86_fp80 @test_f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c) nounwind readnone ssp {
+define x86_fp80 @test_f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c) #0 {
 entry:
-  %call = tail call x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c) nounwind readnone
+  %call = call x86_fp80 @llvm.fma.f80(x86_fp80 %a, x86_fp80 %b, x86_fp80 %c)
   ret x86_fp80 %call
 }
 
-declare float @llvm.fma.f32(float, float, float) nounwind readnone
-declare double @llvm.fma.f64(double, double, double) nounwind readnone
-declare x86_fp80 @llvm.fma.f80(x86_fp80, x86_fp80, x86_fp80) nounwind readnone
+; CHECK-LABEL: test_f32_cst:
+; CHECK-NOT: vfmadd
+define float @test_f32_cst() #0 {
+entry:
+  %call = call float @llvm.fma.f32(float 3.0, float 3.0, float 3.0)
+  ret float %call
+}
+
+declare float @llvm.fma.f32(float, float, float)
+declare double @llvm.fma.f64(double, double, double)
+declare x86_fp80 @llvm.fma.f80(x86_fp80, x86_fp80, x86_fp80)
+
+attributes #0 = { nounwind }