steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-06 13:00:15 +02:00
Code Issues Packages Projects Releases 66 Wiki Activity

OpcodeDispatcher: Optimize pins{b,w,d,q}

Browse Source 
Inserting from a GPR and memory can both be optimized. These are now
optimal

Needs #3088 merged first.
This commit is contained in:
Ryan Houdek committed 2023-09-13 19:53:05 -07:00
1 parent 750d90939d
commit 5e7d793a6a
1 file changed
+10 -11
Download Patch File Download Diff File Expand all files Collapse all files
FEXCore/Source/Interface/Core/OpcodeDispatcher/Vector.cpp
+10 -11
View File
@@ -1501,20 +1501,19 @@ OrderedNode* OpDispatchBuilder::PINSROpImpl(OpcodeArgs, size_t ElementSize,
const X86Tables::DecodedOperand& Imm) {
const auto Size = GetDstSize(Op);
const auto NumElements = Size / ElementSize;
OrderedNode *Src2{};
if (Src2Op.IsGPR()) {
Src2 = LoadSource(GPRClass, Op, Src2Op, Op->Flags, -1);
} else {
// If loading from memory then we only load the element size
Src2 = LoadSource_WithOpSize(GPRClass, Op, Src2Op, ElementSize, Op->Flags, -1);
}
OrderedNode *Src1 = LoadSource_WithOpSize(FPRClass, Op, Src1Op, Size, Op->Flags, -1);
LOGMAN_THROW_A_FMT(Imm.IsLiteral(), "Imm needs to be literal here");
const uint64_t Index = Imm.Data.Literal.Value & (NumElements - 1);
OrderedNode *Src1 = LoadSource_WithOpSize(FPRClass, Op, Src1Op, Size, Op->Flags, -1);
return _VInsGPR(Size, ElementSize, Index, Src1, Src2);
if (Src2Op.IsGPR()) {
// If the source is a GPR then convert directly from the GPR.
auto Src2 = LoadSource_WithOpSize(GPRClass, Op, Src2Op, CTX->GetGPRSize(), Op->Flags, -1);
return _VInsGPR(Size, ElementSize, Index, Src1, Src2);
}
// If loading from memory then we only load the element size
auto Src2 = LoadSource_WithOpSize(GPRClass, Op, Src2Op, ElementSize, Op->Flags, -1, false);
return _VLoadVectorElement(Size, ElementSize, Src1, Index, Src2);
}
template<size_t ElementSize>