maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[AMDGPU] Split out new helper function macToMad in SIFoldOperands. NFC. 

Differential Revision: https://reviews.llvm.org/D94009
This commit is contained in:
Jay Foad 2021-01-04 10:52:22 +00:00
parent 2f8d1e9eb2
commit 4e6054a86c
1 changed files with 22 additions and 29 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
llvm/lib/Target/AMDGPU/SIFoldOperands.cpp
View File

@ -129,6 +129,21 @@ char SIFoldOperands::ID = 0;
char &llvm::SIFoldOperandsID = SIFoldOperands::ID; char &llvm::SIFoldOperandsID = SIFoldOperands::ID;
// Map multiply-accumulate opcode to corresponding multiply-add opcode if any.
static unsigned macToMad(unsigned Opc) {
switch (Opc) {
case AMDGPU::V_MAC_F32_e64:
return AMDGPU::V_MAD_F32;
case AMDGPU::V_MAC_F16_e64:
return AMDGPU::V_MAD_F16;
case AMDGPU::V_FMAC_F32_e64:
return AMDGPU::V_FMA_F32;
case AMDGPU::V_FMAC_F16_e64:
return AMDGPU::V_FMA_F16_gfx9;
}
return AMDGPU::INSTRUCTION_LIST_END;
}
// Wrapper around isInlineConstant that understands special cases when // Wrapper around isInlineConstant that understands special cases when
// instruction types are replaced during operand folding. // instruction types are replaced during operand folding.
static bool isInlineConstantIfFolded(const SIInstrInfo *TII, static bool isInlineConstantIfFolded(const SIInstrInfo *TII,
@ -139,31 +154,18 @@ static bool isInlineConstantIfFolded(const SIInstrInfo *TII,
return true; return true;
unsigned Opc = UseMI.getOpcode(); unsigned Opc = UseMI.getOpcode();
switch (Opc) { unsigned NewOpc = macToMad(Opc);
case AMDGPU::V_MAC_F32_e64: if (NewOpc != AMDGPU::INSTRUCTION_LIST_END) {
case AMDGPU::V_MAC_F16_e64:
case AMDGPU::V_FMAC_F32_e64:
case AMDGPU::V_FMAC_F16_e64: {
// Special case for mac. Since this is replaced with mad when folded into // Special case for mac. Since this is replaced with mad when folded into
// src2, we need to check the legality for the final instruction. // src2, we need to check the legality for the final instruction.
int Src2Idx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::src2); int Src2Idx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::src2);
if (static_cast<int>(OpNo) == Src2Idx) { if (static_cast<int>(OpNo) == Src2Idx) {
bool IsFMA = Opc == AMDGPU::V_FMAC_F32_e64 || const MCInstrDesc &MadDesc = TII->get(NewOpc);
Opc == AMDGPU::V_FMAC_F16_e64;
bool IsF32 = Opc == AMDGPU::V_MAC_F32_e64 ||
Opc == AMDGPU::V_FMAC_F32_e64;
unsigned Opc = IsFMA ?
(IsF32 ? AMDGPU::V_FMA_F32 : AMDGPU::V_FMA_F16_gfx9) :
(IsF32 ? AMDGPU::V_MAD_F32 : AMDGPU::V_MAD_F16);
const MCInstrDesc &MadDesc = TII->get(Opc);
return TII->isInlineConstant(OpToFold, MadDesc.OpInfo[OpNo].OperandType); return TII->isInlineConstant(OpToFold, MadDesc.OpInfo[OpNo].OperandType);
} }
return false;
}
default:
return false;
} }
return false;
} }
// TODO: Add heuristic that the frame index might not fit in the addressing mode // TODO: Add heuristic that the frame index might not fit in the addressing mode
@ -346,17 +348,8 @@ static bool tryAddToFoldList(SmallVectorImpl<FoldCandidate> &FoldList,
if (!TII->isOperandLegal(*MI, OpNo, OpToFold)) { if (!TII->isOperandLegal(*MI, OpNo, OpToFold)) {
// Special case for v_mac_{f16, f32}_e64 if we are trying to fold into src2 // Special case for v_mac_{f16, f32}_e64 if we are trying to fold into src2
unsigned Opc = MI->getOpcode(); unsigned Opc = MI->getOpcode();
if ((Opc == AMDGPU::V_MAC_F32_e64 || Opc == AMDGPU::V_MAC_F16_e64 || unsigned NewOpc = macToMad(Opc);
Opc == AMDGPU::V_FMAC_F32_e64 || Opc == AMDGPU::V_FMAC_F16_e64) && if (NewOpc != AMDGPU::INSTRUCTION_LIST_END) {
(int)OpNo == AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::src2)) {
bool IsFMA = Opc == AMDGPU::V_FMAC_F32_e64 ||
Opc == AMDGPU::V_FMAC_F16_e64;
bool IsF32 = Opc == AMDGPU::V_MAC_F32_e64 ||
Opc == AMDGPU::V_FMAC_F32_e64;
unsigned NewOpc = IsFMA ?
(IsF32 ? AMDGPU::V_FMA_F32 : AMDGPU::V_FMA_F16_gfx9) :
(IsF32 ? AMDGPU::V_MAD_F32 : AMDGPU::V_MAD_F16);
// Check if changing this to a v_mad_{f16, f32} instruction will allow us // Check if changing this to a v_mad_{f16, f32} instruction will allow us
// to fold the operand. // to fold the operand.
MI->setDesc(TII->get(NewOpc)); MI->setDesc(TII->get(NewOpc));