diff options
| author | Jan Vesely <jan.vesely@rutgers.edu> | 2017-01-11 00:12:39 +0000 |
|---|---|---|
| committer | Jan Vesely <jan.vesely@rutgers.edu> | 2017-01-11 00:12:39 +0000 |
| commit | 0d6cb1caaf514cbc2b46096ed1b0a80fa38a6e09 (patch) | |
| tree | c787bdbba36b385d6e8abcde4c121c41176454a7 /llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll | |
| parent | acd63602519a0bc4099bd29357f580a6b5f863f7 (diff) | |
| download | bcm5719-llvm-0d6cb1caaf514cbc2b46096ed1b0a80fa38a6e09.tar.gz bcm5719-llvm-0d6cb1caaf514cbc2b46096ed1b0a80fa38a6e09.zip | |
AMDGPU/EG,CM: Add fp16 conversion instructions
Differential Revision: https://reviews.llvm.org/D28164
llvm-svn: 291622
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll | 22 |
1 files changed, 22 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll b/llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll new file mode 100644 index 00000000000..35e9541692d --- /dev/null +++ b/llvm/test/CodeGen/AMDGPU/fp16_to_fp32.ll @@ -0,0 +1,22 @@ +; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s +; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s +; RUN: llc -march=r600 -mcpu=cypress -verify-machineinstrs < %s | FileCheck -check-prefix=EG -check-prefix=EGCM -check-prefix=FUNC %s +; RUN: llc -march=r600 -mcpu=cayman -verify-machineinstrs < %s | FileCheck -check-prefix=CM -check-prefix=EGCM -check-prefix=FUNC %s + +declare float @llvm.convert.from.fp16.f32(i16) nounwind readnone + +; FUNC-LABEL: {{^}}test_convert_fp16_to_fp32: +; GCN: buffer_load_ushort [[VAL:v[0-9]+]] +; GCN: v_cvt_f32_f16_e32 [[RESULT:v[0-9]+]], [[VAL]] +; GCN: buffer_store_dword [[RESULT]] + +; EG: MEM_RAT_CACHELESS STORE_RAW [[RES:T[0-9]+\.[XYZW]]] +; CM: MEM_RAT_CACHELESS STORE_DWORD [[RES:T[0-9]+\.[XYZW]]] +; EGCM: VTX_READ_16 [[VAL:T[0-9]+\.[XYZW]]] +; EGCM: FLT16_TO_FLT32{{[ *]*}}[[RES]], [[VAL]] +define void @test_convert_fp16_to_fp32(float addrspace(1)* noalias %out, i16 addrspace(1)* noalias %in) nounwind { + %val = load i16, i16 addrspace(1)* %in, align 2 + %cvt = call float @llvm.convert.from.fp16.f32(i16 %val) nounwind readnone + store float %cvt, float addrspace(1)* %out, align 4 + ret void +} |
