RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-09 05:47:13 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

[AMDGPU] Enable code selection using s_mul_hi_u32/s_mul_hi_i32.

Browse Source 
Subscribers: arsenm, kzhuravl, jvesely, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, hiraditya, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D59501

llvm-svn: 356405
This commit is contained in:
Michael Liao 2019-03-18 20:40:09 +00:00
parent 83efed9fe8
commit 4d82c5b7c4
3 changed files with 16 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lib/Target/AMDGPU/SIInstrInfo.cpp
View File

@ -3210,6 +3210,8 @@ unsigned SIInstrInfo::getVALUOp(const MachineInstr &MI) const {
return AMDGPU::V_SUB_I32_e32;
case AMDGPU::S_SUBB_U32: return AMDGPU::V_SUBB_U32_e32;
case AMDGPU::S_MUL_I32: return AMDGPU::V_MUL_LO_I32;
case AMDGPU::S_MUL_HI_U32: return AMDGPU::V_MUL_HI_U32;
case AMDGPU::S_MUL_HI_I32: return AMDGPU::V_MUL_HI_I32;
case AMDGPU::S_AND_B32: return AMDGPU::V_AND_B32_e64;
case AMDGPU::S_OR_B32: return AMDGPU::V_OR_B32_e64;
case AMDGPU::S_XOR_B32: return AMDGPU::V_XOR_B32_e64;
@ -3254,6 +3256,8 @@ unsigned SIInstrInfo::getVALUOp(const MachineInstr &MI) const {
case AMDGPU::S_CBRANCH_SCC0: return AMDGPU::S_CBRANCH_VCCZ;
case AMDGPU::S_CBRANCH_SCC1: return AMDGPU::S_CBRANCH_VCCNZ;
}
llvm_unreachable(
"Unexpected scalar opcode without corresponding vector one!");
}
const TargetRegisterClass *SIInstrInfo::getOpRegClass(const MachineInstr &MI,

8
lib/Target/AMDGPU/SOPInstructions.td
View File

@ -548,8 +548,12 @@ let SubtargetPredicate = isGFX9 in {
def S_LSHL4_ADD_U32 : SOP2_32<"s_lshl4_add_u32">;
} // End Defs = [SCC]
def S_MUL_HI_U32 : SOP2_32<"s_mul_hi_u32">;
def S_MUL_HI_I32 : SOP2_32<"s_mul_hi_i32">;
let isCommutable = 1 in {
def S_MUL_HI_U32 : SOP2_32<"s_mul_hi_u32",
[(set i32:$sdst, (UniformBinFrag<mulhu> SSrc_b32:$src0, SSrc_b32:$src1))]>;
def S_MUL_HI_I32 : SOP2_32<"s_mul_hi_i32",
[(set i32:$sdst, (UniformBinFrag<mulhs> SSrc_b32:$src0, SSrc_b32:$src1))]>;
}
}
//===----------------------------------------------------------------------===//

6
test/CodeGen/AMDGPU/mul.ll
View File

@ -1,5 +1,6 @@
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,SI,FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=GCN,VI,FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=gfx900 -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=FUNC,GFX9 %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mcpu=redwood < %s | FileCheck -allow-deprecated-dag-overlap -check-prefixes=EG,FUNC %s
; mul24 and mad24 are affected
@ -139,6 +140,11 @@ define amdgpu_kernel void @v_mul_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %
; crash with a 'failed to select' error.
; FUNC-LABEL: {{^}}s_mul_i64:
; GFX9-DAG: s_mul_i32
; GFX9-DAG: s_mul_hi_u32
; GFX9-DAG: s_mul_i32
; GFX9-DAG: s_mul_i32
; GFX9: s_endpgm
define amdgpu_kernel void @s_mul_i64(i64 addrspace(1)* %out, i64 %a, i64 %b) nounwind {
%mul = mul i64 %a, %b
store i64 %mul, i64 addrspace(1)* %out, align 8