|
150
|
1 ; RUN: llc -march=amdgcn -verify-machineinstrs < %s
|
|
|
2 ; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s
|
|
|
3 ; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s
|
|
|
4
|
|
|
5 ; This should have the exactly the same output as the test for rint,
|
|
|
6 ; so no need to check anything.
|
|
|
7
|
|
|
8 declare float @llvm.nearbyint.f32(float) #0
|
|
|
9 declare <2 x float> @llvm.nearbyint.v2f32(<2 x float>) #0
|
|
|
10 declare <4 x float> @llvm.nearbyint.v4f32(<4 x float>) #0
|
|
|
11 declare double @llvm.nearbyint.f64(double) #0
|
|
|
12 declare <2 x double> @llvm.nearbyint.v2f64(<2 x double>) #0
|
|
|
13 declare <4 x double> @llvm.nearbyint.v4f64(<4 x double>) #0
|
|
|
14
|
|
|
15
|
|
|
16 define amdgpu_kernel void @fnearbyint_f32(float addrspace(1)* %out, float %in) #1 {
|
|
|
17 entry:
|
|
|
18 %0 = call float @llvm.nearbyint.f32(float %in)
|
|
|
19 store float %0, float addrspace(1)* %out
|
|
|
20 ret void
|
|
|
21 }
|
|
|
22
|
|
|
23 define amdgpu_kernel void @fnearbyint_v2f32(<2 x float> addrspace(1)* %out, <2 x float> %in) #1 {
|
|
|
24 entry:
|
|
|
25 %0 = call <2 x float> @llvm.nearbyint.v2f32(<2 x float> %in)
|
|
|
26 store <2 x float> %0, <2 x float> addrspace(1)* %out
|
|
|
27 ret void
|
|
|
28 }
|
|
|
29
|
|
|
30 define amdgpu_kernel void @fnearbyint_v4f32(<4 x float> addrspace(1)* %out, <4 x float> %in) #1 {
|
|
|
31 entry:
|
|
|
32 %0 = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> %in)
|
|
|
33 store <4 x float> %0, <4 x float> addrspace(1)* %out
|
|
|
34 ret void
|
|
|
35 }
|
|
|
36
|
|
|
37 define amdgpu_kernel void @nearbyint_f64(double addrspace(1)* %out, double %in) {
|
|
|
38 entry:
|
|
|
39 %0 = call double @llvm.nearbyint.f64(double %in)
|
|
|
40 store double %0, double addrspace(1)* %out
|
|
|
41 ret void
|
|
|
42 }
|
|
|
43 define amdgpu_kernel void @nearbyint_v2f64(<2 x double> addrspace(1)* %out, <2 x double> %in) {
|
|
|
44 entry:
|
|
|
45 %0 = call <2 x double> @llvm.nearbyint.v2f64(<2 x double> %in)
|
|
|
46 store <2 x double> %0, <2 x double> addrspace(1)* %out
|
|
|
47 ret void
|
|
|
48 }
|
|
|
49
|
|
|
50 define amdgpu_kernel void @nearbyint_v4f64(<4 x double> addrspace(1)* %out, <4 x double> %in) {
|
|
|
51 entry:
|
|
|
52 %0 = call <4 x double> @llvm.nearbyint.v4f64(<4 x double> %in)
|
|
|
53 store <4 x double> %0, <4 x double> addrspace(1)* %out
|
|
|
54 ret void
|
|
|
55 }
|
|
|
56
|
|
|
57 attributes #0 = { nounwind readonly }
|
|
|
58 attributes #1 = { nounwind }
|
