diff options
| author | Craig Topper <craig.topper@gmail.com> | 2011-12-30 01:49:53 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@gmail.com> | 2011-12-30 01:49:53 +0000 |
| commit | 57d4b3315fb7a84379778c4727594cd7480dbfe1 (patch) | |
| tree | 74526a95f6d729920af006a55e5b7453a27a42f4 /test/CodeGen/X86/fma4-intrinsics-x86_64.ll | |
| parent | 2e95afa04cd1c89de7e4abaeead66cd8e51ec929 (diff) | |
| download | llvm-57d4b3315fb7a84379778c4727594cd7480dbfe1.tar.gz llvm-57d4b3315fb7a84379778c4727594cd7480dbfe1.tar.bz2 llvm-57d4b3315fb7a84379778c4727594cd7480dbfe1.tar.xz | |
Fix load size for FMA4 SS/SD instructions. They need to use f32 and f64 size, but with the special handling to be compatible with the intrinsic expecting a vector. Similar handling is already used elsewhere.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147360 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/X86/fma4-intrinsics-x86_64.ll')
| -rw-r--r-- | test/CodeGen/X86/fma4-intrinsics-x86_64.ll | 28 |
1 files changed, 28 insertions, 0 deletions
diff --git a/test/CodeGen/X86/fma4-intrinsics-x86_64.ll b/test/CodeGen/X86/fma4-intrinsics-x86_64.ll index bd94c134ce..a4b9cc6666 100644 --- a/test/CodeGen/X86/fma4-intrinsics-x86_64.ll +++ b/test/CodeGen/X86/fma4-intrinsics-x86_64.ll @@ -6,6 +6,20 @@ define < 4 x float > @test_x86_fma4_vfmadd_ss(< 4 x float > %a0, < 4 x float > % %res = call < 4 x float > @llvm.x86.fma4.vfmadd.ss(< 4 x float > %a0, < 4 x float > %a1, < 4 x float > %a2) ; <i64> [#uses=1] ret < 4 x float > %res } +define < 4 x float > @test_x86_fma4_vfmadd_ss_load(< 4 x float > %a0, < 4 x float > %a1, float* %a2) { + ; CHECK: vfmaddss (%{{.*}}) + %x = load float *%a2 + %y = insertelement <4 x float> undef, float %x, i32 0 + %res = call < 4 x float > @llvm.x86.fma4.vfmadd.ss(< 4 x float > %a0, < 4 x float > %a1, < 4 x float > %y) ; <i64> [#uses=1] + ret < 4 x float > %res +} +define < 4 x float > @test_x86_fma4_vfmadd_ss_load2(< 4 x float > %a0, float* %a1, < 4 x float > %a2) { + ; CHECK: vfmaddss %{{.*}}, (%{{.*}}) + %x = load float *%a1 + %y = insertelement <4 x float> undef, float %x, i32 0 + %res = call < 4 x float > @llvm.x86.fma4.vfmadd.ss(< 4 x float > %a0, < 4 x float > %y, < 4 x float > %a2) ; <i64> [#uses=1] + ret < 4 x float > %res +} declare < 4 x float > @llvm.x86.fma4.vfmadd.ss(< 4 x float >, < 4 x float >, < 4 x float >) nounwind readnone define < 2 x double > @test_x86_fma4_vfmadd_sd(< 2 x double > %a0, < 2 x double > %a1, < 2 x double > %a2) { @@ -13,6 +27,20 @@ define < 2 x double > @test_x86_fma4_vfmadd_sd(< 2 x double > %a0, < 2 x double %res = call < 2 x double > @llvm.x86.fma4.vfmadd.sd(< 2 x double > %a0, < 2 x double > %a1, < 2 x double > %a2) ; <i64> [#uses=1] ret < 2 x double > %res } +define < 2 x double > @test_x86_fma4_vfmadd_sd_load(< 2 x double > %a0, < 2 x double > %a1, double* %a2) { + ; CHECK: vfmaddsd (%{{.*}}) + %x = load double *%a2 + %y = insertelement <2 x double> undef, double %x, i32 0 + %res = call < 2 x double > @llvm.x86.fma4.vfmadd.sd(< 2 x double > %a0, < 2 x double > %a1, < 2 x double > %y) ; <i64> [#uses=1] + ret < 2 x double > %res +} +define < 2 x double > @test_x86_fma4_vfmadd_sd_load2(< 2 x double > %a0, double* %a1, < 2 x double > %a2) { + ; CHECK: vfmaddsd %{{.*}}, (%{{.*}}) + %x = load double *%a1 + %y = insertelement <2 x double> undef, double %x, i32 0 + %res = call < 2 x double > @llvm.x86.fma4.vfmadd.sd(< 2 x double > %a0, < 2 x double > %y, < 2 x double > %a2) ; <i64> [#uses=1] + ret < 2 x double > %res +} declare < 2 x double > @llvm.x86.fma4.vfmadd.sd(< 2 x double >, < 2 x double >, < 2 x double >) nounwind readnone define < 4 x float > @test_x86_fma4_vfmadd_ps(< 4 x float > %a0, < 4 x float > %a1, < 4 x float > %a2) { |