diff options
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/fpext.f16.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/fpext.f16.ll | 16 |
1 files changed, 10 insertions, 6 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/fpext.f16.ll b/llvm/test/CodeGen/AMDGPU/fpext.f16.ll index 4b03ada925e..03657176c38 100644 --- a/llvm/test/CodeGen/AMDGPU/fpext.f16.ll +++ b/llvm/test/CodeGen/AMDGPU/fpext.f16.ll @@ -1,6 +1,6 @@ -; RUN: llc -march=amdgcn -verify-machineinstrs -enable-unsafe-fp-math < %s | FileCheck -check-prefix=GCN -check-prefix=SI -check-prefix=SIVI %s +; RUN: llc -march=amdgcn -verify-machineinstrs -enable-unsafe-fp-math < %s | FileCheck -check-prefix=GCN -check-prefix=SI -check-prefix=SIVI -check-prefix=SIGFX9 %s ; RUN: llc -march=amdgcn -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs -enable-unsafe-fp-math < %s | FileCheck -check-prefix=GCN -check-prefix=VI -check-prefix=GFX89 %s -; RUN: llc -march=amdgcn -mcpu=gfx901 -mattr=-flat-for-global -verify-machineinstrs -enable-unsafe-fp-math < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -check-prefix=GFX89 %s +; RUN: llc -march=amdgcn -mcpu=gfx901 -mattr=-flat-for-global -verify-machineinstrs -enable-unsafe-fp-math < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -check-prefix=GFX89 -check-prefix=SIGFX9 %s ; GCN-LABEL: {{^}}fpext_f16_to_f32 ; GCN: buffer_load_ushort v[[A_F16:[0-9]+]] @@ -35,12 +35,14 @@ entry: ; GCN-LABEL: {{^}}fpext_v2f16_to_v2f32 ; GCN: buffer_load_dword v[[A_V2_F16:[0-9]+]] -; GFX89-DAG: v_lshrrev_b32_e32 v[[A_F16_1:[0-9]+]], 16, v[[A_V2_F16]] +; GFX9-DAG: v_lshrrev_b32_e32 v[[A_F16_1:[0-9]+]], 16, v[[A_V2_F16]] ; GCN-DAG: v_cvt_f32_f16_e32 v[[R_F32_0:[0-9]+]], v[[A_V2_F16]] ; SI: v_lshrrev_b32_e32 v[[A_F16_1:[0-9]+]], 16, v[[A_V2_F16]] -; GCN: v_cvt_f32_f16_e32 v[[R_F32_1:[0-9]+]], v[[A_F16_1]] +; SIGFX9: v_cvt_f32_f16_e32 v[[R_F32_1:[0-9]+]], v[[A_F16_1]] +; VI: v_cvt_f32_f16_sdwa v[[R_F32_1:[0-9]+]], v[[A_V2_F16]] dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 ; GCN: buffer_store_dwordx2 v{{\[}}[[R_F32_0]]:[[R_F32_1]]{{\]}} ; GCN: s_endpgm + define amdgpu_kernel void @fpext_v2f16_to_v2f32( <2 x float> addrspace(1)* %r, <2 x half> addrspace(1)* %a) #0 { @@ -53,14 +55,16 @@ entry: ; GCN-LABEL: {{^}}fpext_v2f16_to_v2f64 ; GCN: buffer_load_dword -; GCN-DAG: v_lshrrev_b32_e32 -; GCN-DAG: v_cvt_f32_f16_e32 +; SIGFX9-DAG: v_lshrrev_b32_e32 +; SIGFX9-DAG: v_cvt_f32_f16_e32 +; VI: v_cvt_f32_f16_sdwa ; GCN: v_cvt_f32_f16_e32 ; GCN: v_cvt_f64_f32_e32 ; GCN: v_cvt_f64_f32_e32 ; GCN: buffer_store_dwordx4 ; GCN: s_endpgm + define amdgpu_kernel void @fpext_v2f16_to_v2f64( <2 x double> addrspace(1)* %r, <2 x half> addrspace(1)* %a) { |
