Why should I know this?

https://die4taoam.tistory.com/147 computeKnownBits & KnownBits 관련 예제computeKnownBits https://llvm.org/doxygen/ValueTracking_8cpp.html#a903bd19e9d31beff55b22fe86111639e Determine which bits of V are known to be either zero or one and return them. void computeKnownBits (const Value *V, KnownBits &Known, const DataLayout &DL,die4taoam.tistory.com  computeKnownBitsㄴ computeKnownBitsFromContext    ㄴ ..

다음 코드를 통해 computeKnownBitsFromCmp 에서 KnownBit를 추산하는 방식을 알아보자. unsigned BitWidth = Known.getBitWidth(); auto m_V = m_CombineOr(m_Specific(V), m_PtrToIntSameSize(Q.DL, m_Specific(V))); Value *Y; const APInt *Mask, *C; uint64_t ShAmt; switch (Pred) { case ICmpInst::ICMP_EQ: // assume(V = C) if (match(LHS, m_V) && match(RHS, m_APInt(C))) { Known = Known.unionWith(KnownBits::ma..

computeKnownBitsFromCondㄴ computeKnownBitsFromICmpCond void llvm::computeKnownBitsFromICmpCond(const Value *V, ICmpInst *Cmp, KnownBits &Known, const SimplifyQuery &SQ, bool Invert) { ICmpInst::Predicate Pred = Invert ? Cmp->getInversePredicate() : Cmp->getPredicate(); Value *LHS = Cmp->getOperand(0); Value *..

(TBD)

UBFM Syntax UBFM Wd, Wn, #, # ; 32-bit general registers UBFM Xd, Xn, #, # ; 64-bit general registers Usage Unsigned Bitfield Move copies any number of low-order bits from a source register into the same number of adjacent bits at any position in the destination register, with zeros in the upper and lower bits. Important! UBFM is an instruction that copies by specifying a memory range and can be..

https://llvm.org/docs/LangRef.html#load-instruction LLVM Language Reference Manual — LLVM 19.0.0git documentation Compiling with ThinLTO causes the building of a compact summary of the module that is emitted into the bitcode. The summary is emitted into the LLVM assembly and identified in syntax by a caret (’^’). The summary is parsed into a bitcode output, along llvm.org 예제] class BaseLoadStore..

https://github.com/llvm/llvm-project/issues/51483 [AArch64] Generate stp for complex repeating constansts · Issue #51483 · llvm/llvm-project Bugzilla Link 52141 Version trunk OS Windows NT Reporter LLVM Bugzilla Contributor CC @Arnaud-de-Grandmaison-ARM,@DMG862,@smithp35 Extended Description For complex repeating constants like: void fo... github.com 유사한 최적화 제의와 성능 리포트. https://github.com/dotnet..

SmallVector WorkList; WorkList.push_back(RetagReg); while (!WorkList.empty()) { Register UseReg = WorkList.pop_back_val(); for (auto &UseI : MRI->use_instructions(UseReg)) { unsigned Opcode = UseI.getOpcode(); if (Opcode == AArch64::STGi || Opcode == AArch64::ST2Gi || Opcode == AArch64::STZGi || Opcode == AArch64::STZ2Gi || Opcode == AArch64::STGPi || Opcode == AArch64::STGloop || Opcode == AArc..

원문 출처 : https://developer.arm.com/documentation/ddi0602/2023-12/Base-Instructions/STP--Store-Pair-of-Registers-?lang=en#iclass_post_index Documentation – Arm Developer developer.arm.com STP 레지스터 쌍 저장은 기본 레지스터 값과 즉시 오프셋에서 주소를 계산하고 두 개의 레지스터에서 계산된 주소에 32비트 단어 2개 또는 64비트 이중 단어 2개를 저장합니다. 메모리 액세스에 대한 자세한 내용은 로드/저장 주소 지정 모드를 참조하세요. It has encodings from 3 classes: Post-index , Pre-index and Signed of..

* 유관 패치를 작성하게 되면 반영하기 위해 남겨놓는 목적 * * 다른 패치 작성자 분들이 혹시 보게 되면 활용해주세요. * AArch64MIPeepholeOpt.cpp::555 // replaceRegWith changes MIs definition register. Keep it for SSA form until // deleting MI. Only if we made a new destination register. if (DstReg != NewDstReg) { MRI->replaceRegWith(DstReg, NewDstReg); MI.getOperand(0).setReg(DstReg); //