replica
/
hanchenye-llvm-project
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

R600/SI: Add support for llvm.convert.{to|from}.fp16 

llvm-svn: 212676
This commit is contained in:
Matt Arsenault 2014-07-10 03:22:20 +00:00
parent 3e3ddda7a2
commit b0df92577d
3 changed files with 34 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
llvm/lib/Target/R600/SIInstructions.td
View File

@ -1078,8 +1078,12 @@ defm V_CVT_I32_F32 : VOP1_32 <0x00000008, "V_CVT_I32_F32",
[(set i32:$dst, (fp_to_sint f32:$src0))]
>;
defm V_MOV_FED_B32 : VOP1_32 <0x00000009, "V_MOV_FED_B32", []>;
////def V_CVT_F16_F32 : VOP1_F16 <0x0000000a, "V_CVT_F16_F32", []>;
//defm V_CVT_F32_F16 : VOP1_32 <0x0000000b, "V_CVT_F32_F16", []>;
defm V_CVT_F16_F32 : VOP1_32 <0x0000000a, "V_CVT_F16_F32",
[(set i32:$dst, (f32_to_f16 f32:$src0))]
>;
defm V_CVT_F32_F16 : VOP1_32 <0x0000000b, "V_CVT_F32_F16",
[(set f32:$dst, (f16_to_f32 i32:$src0))]
>;
//defm V_CVT_RPI_I32_F32 : VOP1_32 <0x0000000c, "V_CVT_RPI_I32_F32", []>;
//defm V_CVT_FLR_I32_F32 : VOP1_32 <0x0000000d, "V_CVT_FLR_I32_F32", []>;
//defm V_CVT_OFF_F32_I4 : VOP1_32 <0x0000000e, "V_CVT_OFF_F32_I4", []>;

14
llvm/test/CodeGen/R600/fp16_to_fp32.ll Normal file
View File

@ -0,0 +1,14 @@
; RUN: llc -march=r600 -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI %s
declare i16 @llvm.convert.to.fp16(float) nounwind readnone
; SI-LABEL: @test_convert_fp16_to_fp32:
; SI: BUFFER_LOAD_DWORD [[VAL:v[0-9]+]]
; SI: V_CVT_F16_F32_e32 [[RESULT:v[0-9]+]], [[VAL]]
; SI: BUFFER_STORE_SHORT [[RESULT]]
define void @test_convert_fp16_to_fp32(i16 addrspace(1)* noalias %out, float addrspace(1)* noalias %in) nounwind {
%val = load float addrspace(1)* %in, align 4
%cvt = call i16 @llvm.convert.to.fp16(float %val) nounwind readnone
store i16 %cvt, i16 addrspace(1)* %out, align 2
ret void
}

14
llvm/test/CodeGen/R600/fp32_to_fp16.ll Normal file
View File

@ -0,0 +1,14 @@
; RUN: llc -march=r600 -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI %s
declare float @llvm.convert.from.fp16(i16) nounwind readnone
; SI-LABEL: @test_convert_fp16_to_fp32:
; SI: BUFFER_LOAD_USHORT [[VAL:v[0-9]+]]
; SI: V_CVT_F32_F16_e32 [[RESULT:v[0-9]+]], [[VAL]]
; SI: BUFFER_STORE_DWORD [[RESULT]]
define void @test_convert_fp16_to_fp32(float addrspace(1)* noalias %out, i16 addrspace(1)* noalias %in) nounwind {
%val = load i16 addrspace(1)* %in, align 2
%cvt = call float @llvm.convert.from.fp16(i16 %val) nounwind readnone
store float %cvt, float addrspace(1)* %out, align 4
ret void
}