steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-14 07:00:43 +02:00
Code Issues Packages Projects Releases 67 Wiki Activity

Merge pull request #2746 from Sonicadvance1/optimize_maskmov

Browse Source 
OpcodeDispatcher: Optimize MASKMOVDQU and MASKMOVQ
This commit is contained in:
Mai authored and GitHub committed 2023-07-07 22:29:37 -04:00
2 parents 462feec2a6 15f5fe658b
commit 69ea03f0eb
1 file changed
+15 -29
Download Patch File Download Diff File Expand all files Collapse all files
External/FEXCore/Source/Interface/Core/OpcodeDispatcher/Vector.cpp
Vendored
+15 -29
View File
@@ -2257,40 +2257,26 @@ template
void OpDispatchBuilder::XMM_To_MMX_Vector_CVT_Float_To_Int<8, true>(OpcodeArgs);
void OpDispatchBuilder::MASKMOVOp(OpcodeArgs) {
// Until we get correct PHI nodes this is required to be a loop unroll
const auto Size = uint32_t{GetSrcSize(Op)} * 8;
const auto Size = GetSrcSize(Op);
OrderedNode *Src = LoadSource(GPRClass, Op, Op->Src[0], Op->Flags, -1);
OrderedNode *Dest = LoadSource(GPRClass, Op, Op->Dest, Op->Flags, -1);
OrderedNode *MaskSrc = LoadSource(GPRClass, Op, Op->Src[0], Op->Flags, -1);
// Mask only cares about the top bit of each byte
MaskSrc = _VCMPLTZ(Size, 1, MaskSrc);
// Vector that will overwrite byte elements.
OrderedNode *VectorSrc = LoadSource(GPRClass, Op, Op->Dest, Op->Flags, -1);
// RDI source
auto MemDest = LoadGPRRegister(X86State::REG_RDI);
const size_t NumElements = Size / 64;
for (size_t Element = 0; Element < NumElements; ++Element) {
// Extract the current element
auto SrcElement = _VExtractToGPR(GetSrcSize(Op), 8, Src, Element);
auto DestElement = _VExtractToGPR(GetSrcSize(Op), 8, Dest, Element);
// DS prefix by default.
MemDest = AppendSegmentOffset(MemDest, Op->Flags, FEXCore::X86Tables::DecodeFlags::FLAG_DS_PREFIX);
constexpr size_t NumSelectBits = 64 / 8;
for (size_t Select = 0; Select < NumSelectBits; ++Select) {
auto SelectMask = _Bfe(1, 8 * Select + 7, SrcElement);
auto CondJump = _CondJump(SelectMask, {COND_EQ});
auto StoreBlock = CreateNewCodeBlockAfter(GetCurrentBlock());
SetFalseJumpTarget(CondJump, StoreBlock);
SetCurrentCodeBlock(StoreBlock);
{
auto DestByte = _Bfe(8, 8 * Select, DestElement);
auto MemLocation = _Add(MemDest, _Constant(Element * 8 + Select));
// MASKMOVDQU/MASKMOVQ is explicitly weakly-ordered on its store
_StoreMem(GPRClass, 1, MemLocation, DestByte, 1);
}
auto Jump = _Jump();
auto NextJumpTarget = CreateNewCodeBlockAfter(StoreBlock);
SetJumpTarget(Jump, NextJumpTarget);
SetTrueJumpTarget(CondJump, NextJumpTarget);
SetCurrentCodeBlock(NextJumpTarget);
}
}
OrderedNode *XMMReg = _LoadMem(FPRClass, Size, MemDest, 1);
// If the Mask element high bit is set then overwrite the element with the source, else keep the memory variant
XMMReg = _VBSL(Size, MaskSrc, VectorSrc, XMMReg);
_StoreMem(FPRClass, Size, MemDest, XMMReg, 1);
}
void OpDispatchBuilder::VMASKMOVOpImpl(OpcodeArgs, size_t ElementSize, size_t DataSize, bool IsStore,