diff options
Diffstat (limited to 'llvm/test/Analysis/CostModel/AMDGPU/fabs.ll')
| -rw-r--r-- | llvm/test/Analysis/CostModel/AMDGPU/fabs.ll | 18 |
1 files changed, 9 insertions, 9 deletions
diff --git a/llvm/test/Analysis/CostModel/AMDGPU/fabs.ll b/llvm/test/Analysis/CostModel/AMDGPU/fabs.ll index 9c551ec8afe..0d49e2967d2 100644 --- a/llvm/test/Analysis/CostModel/AMDGPU/fabs.ll +++ b/llvm/test/Analysis/CostModel/AMDGPU/fabs.ll @@ -2,7 +2,7 @@ ; CHECK: 'fabs_f32' ; CHECK: estimated cost of 0 for {{.*}} call float @llvm.fabs.f32 -define void @fabs_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr) #0 { %vec = load float, float addrspace(1)* %vaddr %fabs = call float @llvm.fabs.f32(float %vec) #1 store float %fabs, float addrspace(1)* %out @@ -11,7 +11,7 @@ define void @fabs_f32(float addrspace(1)* %out, float addrspace(1)* %vaddr) #0 { ; CHECK: 'fabs_v2f32' ; CHECK: estimated cost of 0 for {{.*}} call <2 x float> @llvm.fabs.v2f32 -define void @fabs_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1)* %vaddr) #0 { %vec = load <2 x float>, <2 x float> addrspace(1)* %vaddr %fabs = call <2 x float> @llvm.fabs.v2f32(<2 x float> %vec) #1 store <2 x float> %fabs, <2 x float> addrspace(1)* %out @@ -20,7 +20,7 @@ define void @fabs_v2f32(<2 x float> addrspace(1)* %out, <2 x float> addrspace(1) ; CHECK: 'fabs_v3f32' ; CHECK: estimated cost of 0 for {{.*}} call <3 x float> @llvm.fabs.v3f32 -define void @fabs_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1)* %vaddr) #0 { %vec = load <3 x float>, <3 x float> addrspace(1)* %vaddr %fabs = call <3 x float> @llvm.fabs.v3f32(<3 x float> %vec) #1 store <3 x float> %fabs, <3 x float> addrspace(1)* %out @@ -29,7 +29,7 @@ define void @fabs_v3f32(<3 x float> addrspace(1)* %out, <3 x float> addrspace(1) ; CHECK: 'fabs_f64' ; CHECK: estimated cost of 0 for {{.*}} call double @llvm.fabs.f64 -define void @fabs_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr) #0 { %vec = load double, double addrspace(1)* %vaddr %fabs = call double @llvm.fabs.f64(double %vec) #1 store double %fabs, double addrspace(1)* %out @@ -38,7 +38,7 @@ define void @fabs_f64(double addrspace(1)* %out, double addrspace(1)* %vaddr) #0 ; CHECK: 'fabs_v2f64' ; CHECK: estimated cost of 0 for {{.*}} call <2 x double> @llvm.fabs.v2f64 -define void @fabs_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %vaddr) #0 { %vec = load <2 x double>, <2 x double> addrspace(1)* %vaddr %fabs = call <2 x double> @llvm.fabs.v2f64(<2 x double> %vec) #1 store <2 x double> %fabs, <2 x double> addrspace(1)* %out @@ -47,7 +47,7 @@ define void @fabs_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace( ; CHECK: 'fabs_v3f64' ; CHECK: estimated cost of 0 for {{.*}} call <3 x double> @llvm.fabs.v3f64 -define void @fabs_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace(1)* %vaddr) #0 { %vec = load <3 x double>, <3 x double> addrspace(1)* %vaddr %fabs = call <3 x double> @llvm.fabs.v3f64(<3 x double> %vec) #1 store <3 x double> %fabs, <3 x double> addrspace(1)* %out @@ -56,7 +56,7 @@ define void @fabs_v3f64(<3 x double> addrspace(1)* %out, <3 x double> addrspace( ; CHECK: 'fabs_f16' ; CHECK: estimated cost of 0 for {{.*}} call half @llvm.fabs.f16 -define void @fabs_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr) #0 { %vec = load half, half addrspace(1)* %vaddr %fabs = call half @llvm.fabs.f16(half %vec) #1 store half %fabs, half addrspace(1)* %out @@ -65,7 +65,7 @@ define void @fabs_f16(half addrspace(1)* %out, half addrspace(1)* %vaddr) #0 { ; CHECK: 'fabs_v2f16' ; CHECK: estimated cost of 0 for {{.*}} call <2 x half> @llvm.fabs.v2f16 -define void @fabs_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)* %vaddr) #0 { %vec = load <2 x half>, <2 x half> addrspace(1)* %vaddr %fabs = call <2 x half> @llvm.fabs.v2f16(<2 x half> %vec) #1 store <2 x half> %fabs, <2 x half> addrspace(1)* %out @@ -74,7 +74,7 @@ define void @fabs_v2f16(<2 x half> addrspace(1)* %out, <2 x half> addrspace(1)* ; CHECK: 'fabs_v3f16' ; CHECK: estimated cost of 0 for {{.*}} call <3 x half> @llvm.fabs.v3f16 -define void @fabs_v3f16(<3 x half> addrspace(1)* %out, <3 x half> addrspace(1)* %vaddr) #0 { +define amdgpu_kernel void @fabs_v3f16(<3 x half> addrspace(1)* %out, <3 x half> addrspace(1)* %vaddr) #0 { %vec = load <3 x half>, <3 x half> addrspace(1)* %vaddr %fabs = call <3 x half> @llvm.fabs.v3f16(<3 x half> %vec) #1 store <3 x half> %fabs, <3 x half> addrspace(1)* %out |
