The Week Of Monday 5 October 2026 Archives by thread
Starting: Mon Oct 5 00:03:03 PDT 2026
Ending: Sun Oct 11 23:57:27 PDT 2026
Messages: 7551
- [llvm] [X86] Use INTR_TYPE_1OP_MASK_SAE for avx10_mask_vcvttps2ibs512. (PR #228988)
Sergei Barannikov via llvm-commits
- [llvm] [MC] Diagnose negative CFI address deltas (PR #228663)
Alexis Engelke via llvm-commits
- [llvm] [AMDGPU] Preserve live SCC when shrinking disjoint S_OR_B32 to S_ADDK_I32 (PR #228025)
Teja Alaghari via llvm-commits
- [llvm] [SPIRV] Only emit MaxByteOffset decoration for SPIR-V 1.1 or later (PR #228973)
Arseniy Obolenskiy via llvm-commits
- [llvm] [llvm] Fix string and Twine concatenation for C++26 P2591R5 (PR #227959)
James Henderson via llvm-commits
- [llvm] [AMDGPU][SwLowerLDS] Rebase flat to LDS addrspacecasts to buffer offsets (PR #227985)
via llvm-commits
- [llvm] [Attributor] Only follow addrspacecasts in AAPointerInfo (PR #225290)
Nikita Popov via llvm-commits
- [llvm] [GVN] create replaceWithEquivalentCmp function (NFC) (PR #228579)
Antonio Frighetto via llvm-commits
- [llvm] [RISCV] Fixed redundant add instructions generated during the fold of vscale add in rvv. (PR #225266)
Kamlesh Kumar via llvm-commits
- [llvm] [GVN] Refactor code into `replaceWithEquivalentCmp` function (NFC) (PR #228579)
Antonio Frighetto via llvm-commits
- [llvm] [SPIRV] Respect source alignment when lowering memory copies (PR #228859)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AArch64] Mark masked load/store data types that get legalized to scalars illegal (PR #228122)
Mattéo Rizza Murgier via llvm-commits
- [llvm] [ValueTracking] Teach isGuaranteedNotToBeUndefOrPoison about ZExt (PR #228493)
Nikita Popov via llvm-commits
- [llvm] [AMDGPU] Clamp promote alloca LDS limit to addressable LDS size (PR #228987)
Arseniy Obolenskiy via llvm-commits
- [llvm] address unused-vars in the no-assert build for PR #213007 (PR #228986)
LLVM Continuous Integration via llvm-commits
- [llvm] [AggressiveInstCombine] Extend MulHigh fold for `or disjoint` (PR #228897)
Nikita Popov via llvm-commits
- [llvm] [AMDGPU] Fix LDS access via flat pointer argument in amdgpu-sw-lower-lds (PR #209842)
via llvm-commits
- [llvm] c77c788 - [X86] Don't pass SAE operand to some nodes. (#228938)
via llvm-commits
- [llvm] [X86] Don't pass SAE operand to some nodes. (PR #228938)
Craig Topper via llvm-commits
- [llvm] [BasicAA] Remove TODO in test (NFC) (PR #228992)
Nikita Popov via llvm-commits
- [llvm] [Analysis] Remove unused declarations (NFC) (PR #228993)
Kazu Hirata via llvm-commits
- [llvm] [AMDGPU][SROA] Expand cast chain handling to floating point types (PR #228113)
Steffen Larsen via llvm-commits
- [llvm] [VectorCombine] Fold insertelement chains of scalar parts to a bitcast and shuffle (PR #226224)
Tim Besard via llvm-commits
- [llvm] [DAG] Add basic deinterleave/interleave(poison) -> poison combines. (PR #228858)
Kamlesh Kumar via llvm-commits
- [llvm] [TailCallElim] Do not mark a call tail when it is handed the frame (PR #218797)
Antonio Frighetto via llvm-commits
- [llvm] [ADT] Clean up DenseMap.h after removing CRTP (NFC) (PR #228981)
Kazu Hirata via llvm-commits
- [llvm] aaa9228 - [TailCallElim] Do not mark a call tail when it is handed the frame (#218797)
via llvm-commits
- [llvm] [AArch64][GlobalISel] Add deinterleaved store combine. (PR #228514)
Kamlesh Kumar via llvm-commits
- [llvm] [KnownBits] Improve minimum denominator bounds for sdiv (PR #227264)
via llvm-commits
- [llvm] [LV] Emit truncate for widen-induction in VPlan during construction. (PR #225991)
Elvis Wang via llvm-commits
- [llvm] [BOLT][DWARF] Fix unit layout with forward DW_FORM_ref_udata references (PR #226076)
Tim Besard via llvm-commits
- [llvm] c30b45e - [NFC][IR] Use getDataLayout member function for Function and GlobalValue (#228964)
via llvm-commits
- [llvm] [NFC][IR] Use getDataLayout member function for Function and GlobalValue (PR #228964)
Nikita Popov via llvm-commits
- [llvm] [GVN] Allow non local MemoryDef tries to run (PR #228836)
Antonio Frighetto via llvm-commits
- [llvm] [NFC][IR] Use getDataLayout member function for Instruction and BasicBlock (PR #228989)
Nikita Popov via llvm-commits
- [llvm] 451e086 - [NFC][IR] Use getDataLayout member function for Instruction and BasicBlock (#228989)
via llvm-commits
- [llvm] [RISCV] Refactor copy from X0 recognition in RISCVRedundantCopyElimination. (PR #228661)
Sudharsan Veeravalli via llvm-commits
- [clang] [llvm] [Clang][AIX] Add -mloadtime-comment-vars flag to preserve identifying variables (PR #187986)
Tony Varghese via llvm-commits
- [llvm] [X86] Use CF from SHR for d & 1 when d >> 1 is also computed (PR #228181)
via llvm-commits
- [llvm] [VectorCombine] Fold interleave and widen chained operations (PR #224005)
Kamlesh Kumar via llvm-commits
- [llvm] [llvm-c][ocaml] Deprecate/replace constexpr GEP APIs (PR #228998)
Nikita Popov via llvm-commits
- [llvm] [Mips] Fold FP boundary comparisons to CLASS on MIPS R6 (PR #229000)
Julian Zhu via llvm-commits
- [llvm] [NVPTX] Set default value of nvptx-allow-ftz-atomics to true (PR #206154)
Ilya Biryukov via llvm-commits
- [llvm] [SeparateConstOffsetFromGEP] Use std::optional over 0 sentinel for extraction failure (PR #228538)
Antonio Frighetto via llvm-commits
- [llvm] [AMDGPU] Fix overlapping insert crash during rewrite-agpr-copy-mfma (PR #205962)
Lukas Sommer via llvm-commits
- [llvm] [X86][SSE] Improve ISD::FP_TO_*INT_SAT vector lowering (PR #199416)
via llvm-commits
- [llvm] [X86] Don't treat vectors with elements wider than 64 bits as immediate-shiftable (PR #228840)
Simon Pilgrim via llvm-commits
- [llvm] [VPlan] Constant-fold ActiveLaneMask intrinsics (PR #208852)
David Sherwood via llvm-commits
- [llvm] [X86][SchedModel] Add masked gather/scatter overrides to Znver4Model (PR #212997)
Sumukh J Bharadwaj via llvm-commits
- [llvm] 2ab8bb0 - [BasicAA] Remove TODO in test (NFC) (#228992)
via llvm-commits
- [llvm] [PGO] Fix PGO instrumentation for selects with poison condition (PR #229002)
Nikita Popov via llvm-commits
- [clang] [llvm] [TimeProfiler][Clang] Add zstd-compressed time-trace output support (PR #228627)
Chandler Carruth via llvm-commits
- [llvm] [GlobalISel] [AArch64] Generate MVN for shift by (N - X) in SelectShiftMask (PR #226273)
Deepak Shirke via llvm-commits
- [llvm] [InstCombine] Ensure i1 smulh(x,1) fold to zero (PR #229004)
Simon Pilgrim via llvm-commits
- [llvm] [RISCV] Add Xqcisls loads to getJumpTableIndex (PR #229006)
Sudharsan Veeravalli via llvm-commits
- [clang] [llvm] [RISC-V] Disable save-restore with shadow stack (PR #225610)
Nemanja Ivanovic via llvm-commits
- [llvm] [GlobalISel][AArch64] Lower G_BITCAST vectors with non-integer element ratio (PR #215154)
Kacper Doga via llvm-commits
- [llvm] [InstCombine] Add generic smulh/umulh identity folds (PR #227335)
Simon Pilgrim via llvm-commits
- [llvm] [EarlyIfConversion] Limit critical-path growth for strongly biased branches (PR #228709)
Jonathan Cohen via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
James Henderson via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
Tomohiro Kashiwada via llvm-commits
- [llvm] [llvm-strings] Add --encoding={s,S,utf8} option (PR #221794)
Harald van Dijk via llvm-commits
- [llvm] [AMDGPU][docs] define cluster scope and multicast loads (gfx1250) (PR #225310)
Pierre van Houtryve via llvm-commits
- [llvm] [SeparateConstOffsetFromGEP] Fix sext/add overflow checks (PR #228562)
Antonio Frighetto via llvm-commits
- [llvm] AMDGPU: Stop setting kill flags before FinalizeISel (PR #229009)
Matt Arsenault via llvm-commits
- [llvm] [GlobalISel] Translate byte to ptr bitcasts to `G_INTTOPTR`/`G_PTRTOINT` (PR #226541)
Pedro Lobo via llvm-commits
- [llvm] [APFloat] Allow lossless conversions to PPCDoubleDouble (PR #228810)
Tim Gymnich via llvm-commits
- [llvm] VE: Fix ill-typed setjmp result in emitEHSjLjSetJmp (PR #221549)
Matt Arsenault via llvm-commits
- [llvm] [CGP] Simple loop strength reduction for vector values (PR #226197)
Sander de Smalen via llvm-commits
- [llvm] [LV] Make partial-reduction naming consistent (NFC) (PR #222377)
Sander de Smalen via llvm-commits
- [llvm] 253db65 - VE: Fix ill-typed setjmp result in emitEHSjLjSetJmp (#221549)
via llvm-commits
- [llvm] VE: Compute live-ins after splitting for EXTEND_STACK expansion (PR #221553)
Matt Arsenault via llvm-commits
- [llvm] [X86] Let RegAllocFast lower tied operands (PR #228968)
Simon Pilgrim via llvm-commits
- [llvm] [AArch64][LSR] Prefer pointer IVs for SVE accesses that need splitting (PR #228510)
Benjamin Maxwell via llvm-commits
- [clang] [llvm] [AArch64] Add CMH hints to store_with_hint intrinsic (PR #227326)
via llvm-commits
- [llvm] [AMDGPU] Check bundled instructions in canUsePressureDiffs (PR #227620)
Pierre van Houtryve via llvm-commits
- [llvm] [SLP]Vectorize aggregates of bit packs as a single vector (PR #228955)
Simon Pilgrim via llvm-commits
- [llvm] [RISCV] Ensure Shadow Stack checks correct RA (PR #226309)
Nemanja Ivanovic via llvm-commits
- [llvm] VE: Remove broken nested call frame around dynamic stack allocation (PR #229011)
Matt Arsenault via llvm-commits
- [llvm] AMDGPU/GlobalISel: Legalize bf16 fsin and fcos (PR #227075)
Petar Avramovic via llvm-commits
- [llvm] [Dexter] Add timeout and finishing guard for startup (PR #228636)
Stephen Tozer via llvm-commits
- [llvm] [X86] Fix bool vector extension for non-power-of-2 sizes (PR #228921)
Simon Pilgrim via llvm-commits
- [clang] [llvm] [clang][NVPTX] Add overloaded fmul intrinsics (PR #224546)
Srinivasa Ravi via llvm-commits
- [llvm] AMDGPU/GlobalISel: Legalize bf16 fdiv (PR #227136)
Petar Avramovic via llvm-commits
- [llvm] [llvm-objcopy] Preserve symbols with st_shndx in [SHN_LOOS, SHN_HIOS] (PR #228665)
James Henderson via llvm-commits
- [llvm] [MachineCombiner] Don't count debug instructions in block size (PR #225765)
Bill Wendling via llvm-commits
- [llvm] [AMDGPU][GlobalISel] Keep typed LLTs in RegBankLegalize combines (PR #227476)
Petar Avramovic via llvm-commits
- [llvm] VE: Stop setting kill flags on virtual registers before FinalizeISel (PR #229012)
Matt Arsenault via llvm-commits
- [llvm] [BOLT][AArch64] Remove thunk estimation and warn on margin exhaustion (PR #228069)
Alexandros Lamprineas via llvm-commits
- [llvm] [AMDGPU] Skip the iterative atomic scan on native LDS atomics (PR #227222)
Carl Ritson via llvm-commits
- [llvm] RISCV: Use hasOneNonDBGUse instead of kill flag for cascaded select (PR #228940)
Matt Arsenault via llvm-commits
- [llvm] 0283f49 - [InstCombine] Ensure i1 smulh(x, 1) fold to zero (#229004)
via llvm-commits
- [llvm] [X86] Promote approximate v8f16 FMA to v8f32 (PR #228900)
Simon Pilgrim via llvm-commits
- [llvm] [LoopVectorize] Drop SCEV wrap predicate for cast-based inductions with all-narrowing uses (PR #228847)
David Sherwood via llvm-commits
- [llvm] RISCV: Stop setting kill flags on virtual registers before FinalizeISel (PR #229013)
Matt Arsenault via llvm-commits
- [llvm] [DAG] FoldConstantArithmetic - add support for ISD::MULO/ADDO/SUBO constant folding (PR #228877)
Simon Pilgrim via llvm-commits
- [llvm] [Bazel] Fixes a803686 (PR #229014)
via llvm-commits
- [llvm] [offload][omp] Move reading _kernel_environment to libomptarget (PR #222606)
Alex Duran via llvm-commits
- [llvm] [llvm-dwarfdump] Added .debug_abbrev section for --output-style format (PR #227195)
James Henderson via llvm-commits
- [llvm] [Attributor] Do not use a recursive callee's values as the call result (PR #229015)
via llvm-commits
- [llvm] PowerPC/GlobalISel: Stop setting kill flags on selected instructions (PR #229016)
Matt Arsenault via llvm-commits
- [llvm] [NFC][AMDGPU] Precommit test for byte lane extension after vector split (PR #228423)
Dmitry Sidorov via llvm-commits
- [llvm] [offload][OpenMP] Set StaticBlockMemSize in kernel launch info (PR #228756)
Alex Duran via llvm-commits
- [llvm] [LAA] Avoid duplicate runtime-check group members (PR #226829)
Florian Hahn via llvm-commits
- [llvm] Mips: Stop setting kill flags on virtual registers before FinalizeISel (PR #229018)
Matt Arsenault via llvm-commits
- [llvm] [LoopVectorize] Fix double-application of FindIV reduction expression in epilogue (PR #219362)
Florian Hahn via llvm-commits
- [llvm] [X86] Fold FMADDSUB into VFMULC for fp16 complex multiply (PR #227698)
Rohan Shenoy via llvm-commits
- [llvm] [llvm-profgen] Add executable range-end analysis (PR #224993)
Sergey Shcherbinin via llvm-commits
- [clang] [llvm] [SPIRV] Add size optimisation for AMDGCN flavoured SPIR-V (PR #228188)
Marcos Maronas via llvm-commits
- [llvm] [SDPatternMatch] Add m_Node and m_SpecificOpc that take opcode from template argument (PR #228626)
Simon Pilgrim via llvm-commits
- [llvm] [KnownFPClass] Refactor NaN sign encoding for KnownFPClass (PR #220606)
Tim Gymnich via llvm-commits
- [llvm] AMDGPU/GlobalISel: Legalize bf16 g_fcmp (PR #227909)
Petar Avramovic via llvm-commits
- [llvm] [LV] Add cross-part load overlap heuristic for interleaving (PR #214500)
Sergey Shcherbinin via llvm-commits
- [llvm] LoongArch: Stop setting kill flags on virtual registers before FinalizeISel (PR #229019)
Matt Arsenault via llvm-commits
- [llvm] [LV] Add tests for uniform inner loops in outer-loop path (NFC) (PR #227415)
Florian Hahn via llvm-commits
- [flang] [llvm] [flang] Support static-unit array slices in FIR LoopVersioning (PR #222723)
Sergey Shcherbinin via llvm-commits
- [llvm] [InstCombine] Fold icmp ugt (sdiv exact X, C2), C into a range check (PR #221577)
armandeep singh via llvm-commits
- [llvm] ARM: Stop setting kill flags on virtual registers before FinalizeISel (PR #229020)
Matt Arsenault via llvm-commits
- [llvm] [GlobalISel] Add GIHasOneUse builtin for combiner match patterns (PR #221804)
Pierre van Houtryve via llvm-commits
- [llvm] AArch64: Stop setting kill flags on virtual registers before FinalizeISel (PR #229021)
Matt Arsenault via llvm-commits
- [llvm] [LangRef] Clarify invariant.group semantics (PR #224222)
Antonio Frighetto via llvm-commits
- [llvm] [Support] Switch _Pragma to __pragma w/MSVC; NFC (PR #228545)
Hans Wennborg via llvm-commits
- [llvm] X86: Stop setting kill flags on virtual registers before FinalizeISel (PR #229022)
Matt Arsenault via llvm-commits
- [llvm] [SPIR-V] Lower get_fence builtin to OpGenericPtrMemSemantics (PR #229023)
Arseniy Obolenskiy via llvm-commits
- [llvm] [Remarks] Fix use-after-free of block scalar values in the YAML remark parser (PR #228072)
Florian Hahn via llvm-commits
- [llvm] [CodeGenPrepare] Recursively sink expensive operands when converting select to branch (PR #197869)
via llvm-commits
- [llvm] [opt-viewer] Fix demangler usage with multiprocessing (PR #228388)
Florian Hahn via llvm-commits
- [llvm] fb89552 - [GlobalISel] Translate byte to ptr bitcasts to `G_INTTOPTR`/`G_PTRTOINT` (#226541)
via llvm-commits
- [llvm] [LoopVectorize] Support vectorization of compressing patterns (PR #214491)
Benjamin Maxwell via llvm-commits
- [llvm] [llvm-objcopy][COFF] Keep COMDAT section definition symbols when stripping (PR #228783)
Martin Storsjö via llvm-commits
- [llvm] X86: Preserve the dead flag clobber when expanding dynamic allocas (PR #227702)
Simon Pilgrim via llvm-commits
- [llvm] [LV] Fix double-application of FindIV reduction expression in epilogue (PR #219362)
Florian Hahn via llvm-commits
- [llvm] [BOLT][AArch64] Don't patch short-range PC-relative refs in un-emitte… (PR #221344)
Alexandros Lamprineas via llvm-commits
- [llvm] [X86] Declare memory effects for the AMX intrinsics (PR #229025)
Jarrod Petz via llvm-commits
- [llvm] CodeGen: Prefer getting the Triple from the Module (PR #228620)
Marcos Maronas via llvm-commits
- [llvm] f96e385 - VE: Compute live-ins after splitting for EXTEND_STACK expansion (#221553)
via llvm-commits
- [llvm] VE: Use splitAt in expandExtendStackPseudo (PR #221554)
Matt Arsenault via llvm-commits
- [llvm] [LV][NFC] Fix "vectorization is possible but not beneficial" message (PR #228024)
Florian Hahn via llvm-commits
- [llvm] 07cd029 - [LV] Make partial-reduction naming consistent (NFC) (#222377)
via llvm-commits
- [llvm] CodeGen: Prefer getting the Triple from the Module (PR #228682)
Marcos Maronas via llvm-commits
- [llvm] [Offload][L0][NFC] Remove unused zexKernelGetArgumentSize dispatcher (PR #227589)
via llvm-commits
- [llvm] [X86] Fix reversed condition in ADC/SBB fold of COND_A compares (PR #228693)
Simon Pilgrim via llvm-commits
- [llvm] [X86] Swap CMP operands for COND_A in ADC/SBB fold (PR #228710)
Simon Pilgrim via llvm-commits
- [llvm] [SPIR-V] Avoid dead OpLogicalOr with false in legalizeIsFPClass (PR #216639)
Marcos Maronas via llvm-commits
- [llvm] [LoopVectorize][VPlan] Support first-only-fault loads in early exits (PR #229027)
David Del Río via llvm-commits
- [llvm] 73ad673 - VE: Stop setting kill flags on virtual registers before FinalizeISel (#229012)
via llvm-commits
- [llvm] [AArch64][Win] Account for the fixed object area when classifying CSR (PR #222111)
Benjamin Maxwell via llvm-commits
- [llvm] X86: don't fallback to X86-32 calling convention in 64bit mode. (PR #225820)
Jacob Lifshay via llvm-commits
- [llvm] [VPlan] Compute scalar cost based on VPlan0 instead of legacy CM. (PR #196845)
Florian Hahn via llvm-commits
- [llvm] 795e750 - PowerPC/GlobalISel: Stop setting kill flags on selected instructions (#229016)
via llvm-commits
- [llvm] X86: Avoid keeping EFLAGS live for an unused cmpxchg success value (PR #227506)
Simon Pilgrim via llvm-commits
- [llvm] 5b2879a - [AMDGPU] Skip the iterative atomic scan on native LDS atomics (#227222)
via llvm-commits
- [llvm] [VPlan] Model diff checks checks as VPlan recipes. (PR #229028)
Florian Hahn via llvm-commits
- [llvm] [SPIR-V] Fix DebugScope resolving DILexicalBlockFile to DebugCompilationUnit (PR #228259)
Manuel Carrasco via llvm-commits
- [llvm] [CostModel][X86] Add basic AVX512 Intrinsic::masked_compresstore cost tables (PR #226240)
Rohan Shenoy via llvm-commits
- [llvm] [Support] allow placement new with extended alignment (PR #228507)
Rose Hudson via llvm-commits
- [llvm] [AArch64] Transform lsr + cbz/cbnz into cmp + bcc in loops (PR #228005)
via llvm-commits
- [llvm] [X86] Don't allocate unused padding in front of realigned locals (PR #227495)
Timur Golubovich via llvm-commits
- [llvm] e1ab1b1 - [X86] Fold FMADDSUB into VFMULC for fp16 complex multiply (#227698)
via llvm-commits
- [llvm] [AArch64] Prefer tst + bcc over lsr + cbz/cbnz (PR #228005)
via llvm-commits
- [clang] [llvm] [RISCV] Add Tenstorrent Ascalon XG (PR #229030)
Ramkumar Ramachandra via llvm-commits
- [llvm] 042302b - [AArch64] Add CMH hints to store_with_hint intrinsic (#227326)
via llvm-commits
- [llvm] 5b07ba7 - [llvm-c][ocaml] Deprecate/replace constexpr GEP APIs (#228998)
via llvm-commits
- [llvm] [ConstraintElim] Bound IVs by compares of the phi in header/latch. (PR #226297)
Florian Hahn via llvm-commits
- [llvm] [LV] Allow for vectorizing parallel def-use chains (PR #225405)
Nashe Mncube via llvm-commits
- [llvm] [SCEV] Preserve nuw on scaled recurrence start values. (PR #227458)
Florian Hahn via llvm-commits
- [polly] [Polly] Distribute the innermost point loop of tile over its statements (PR #225844)
Michael Kruse via llvm-commits
- [llvm] [DebugInfo] Preserve value semantics when salvaging truncations (PR #229033)
Kamil Monicz via llvm-commits
- [llvm] [AMDGPU] Track DS prefetch flush points on single-path loop. (PR #227274)
Vikash Gupta via llvm-commits
- [llvm] [X86][CostModel] Add per-shape gather/scatter cost tables for AMD znver4+ (PR #199488)
Sumukh J Bharadwaj via llvm-commits
- [llvm] [NFC][AMDGPU] Precommit test for MFMA accumulator register hint (PR #228531)
Romanov Vlad via llvm-commits
- [llvm] [InstCombine] Limit recursion depth when evaluating in a different type (PR #217879)
Maksim Shelegov via llvm-commits
- [llvm] [SCEV] Allow replacing disjoint ors when reusing instructions (PR #224246)
Maksim Shelegov via llvm-commits
- [lld] [lld][COFF] Add per-target runtime directories to library search path (PR #227744)
Jacek Caban via llvm-commits
- [llvm] [SimplifyCFG] Optimize switches on constant globals loads (PR #229035)
Ömer Sinan Ağacan via llvm-commits
- [llvm] [KnownFPClass] Correct denormal handling for `KnownFPClass::roundToIntegral` (PR #219700)
Tim Gymnich via llvm-commits
- [llvm] 9683a9b - [X86] Fix reversed condition in ADC/SBB fold of COND_A compares (#228693)
via llvm-commits
- [clang] [llvm] [offload-arch] Move GPU detection into a library and unit tests to unittests/tools (PR #226552)
Nikita Kornev via llvm-commits
- [llvm] [LV] Add tests for reverse early-exit loops in deref regions (PR #228062)
Aleksandr Popov via llvm-commits
- [llvm] [SCEV] Exit early for all-canonical ops in computeAndSetCanonical (NFC) (PR #229036)
Florian Hahn via llvm-commits
- [llvm] [libsycl] Introduce Kernel name based cache (PR #225865)
via llvm-commits
- [llvm] [SLP]Keep dbg_value of the lanes replaced by extractelement (PR #228937)
Alexey Bataev via llvm-commits
- [llvm] [SLP]Fix masked gather GEPs cost, skip GEP index seeds (PR #228918)
Alexey Bataev via llvm-commits
- [llvm] [SLP]Vectorize consecutive loads with undef lanes as a wide load (PR #228872)
Alexey Bataev via llvm-commits
- [llvm] [Reassociate]Reject fma pairs with a shared addend (PR #228751)
Alexey Bataev via llvm-commits
- [llvm] [MachineCombiner]Reassociate FP accumulation chains to expose ILP (PR #222676)
Alexey Bataev via llvm-commits
- [llvm] [SLP] Enable full non-power-of-2 vectorization by default (PR #199450)
Alexey Bataev via llvm-commits
- [llvm] [SPIRV] LLVM IR DIScope nodes are now emitted after their dependencies. (PR #227361)
Manuel Carrasco via llvm-commits
- [llvm] [AArch64] Fix incorrect frame indexes for ZPR callee-saved pair spills (PR #228368)
Kieran B via llvm-commits
- [llvm] [NFC][SLP] Precommit tests for the phantom load saving (PR #228901)
Alexey Bataev via llvm-commits
- [llvm] [SPIRV] Lower OpenCL sub_group_barrier to OpControlBarrier (PR #228963)
Marcos Maronas via llvm-commits
- [clang] [llvm] [AArch64][SME] Allow more inlining when SME attributes are incompatible. (PR #223393)
Sander de Smalen via llvm-commits
- [llvm] [LV] Allow out-of-loop backedge users for min/max value of argmin/argmax. (PR #220351)
Nashe Mncube via llvm-commits
- [llvm] [SimplifyCFG] Preserve AA metadata when speculating stores (PR #227034)
Hari Limaye via llvm-commits
- [llvm] [AMDGPU] Add heuristic to set CarriedLatencies (PR #224057)
Romanov Vlad via llvm-commits
- [llvm] [KnownFPClass] Fix input/output subnormal handling for `propagateXorSign` (PR #219837)
Tim Gymnich via llvm-commits
- [llvm] [AMDGPU][CoExecSched] Add free-proximity scheduling heuristic (PR #225334)
Romanov Vlad via llvm-commits
- [llvm] [AArch64] Add sext(shl x, imm)/sext(sra x, imm) selection patterns (PR #226065)
Hari Limaye via llvm-commits
- [llvm] [LV] Narrow truncated inductions in a VPlan transform (PR #220730)
Alexander Kornienko via llvm-commits
- [llvm] PowerPC: Mark TLS call pseudo clobbers dead at build time (PR #227050)
Matt Arsenault via llvm-commits
- [llvm] 267acde - [VectorCombine] Fold interleave and widen chained operations (#224005)
via llvm-commits
- [llvm] [LiveDebugVariables] Stop holding SlotIndexes for erased instructions (PR #224467)
Jeremy Morse via llvm-commits
- [llvm] [MoveAutoInit] Don't move auto-init instructions into EH pad blocks (PR #222100)
Jeremy Morse via llvm-commits
- [llvm] fbf5869 - [MoveAutoInit] Don't move auto-init instructions into EH pad blocks (#222100)
via llvm-commits
- [llvm] [CodeView] Support LF_CLASS2 and friends (PR #228761)
via llvm-commits
- [llvm] [TableGen] Avoid crashing on a missing instruction pattern result (PR #228075)
Jonathan Thackray via llvm-commits
- [llvm] [TableGen] Assign LaneBitmask bits in a less surprising order (PR #225809)
Jay Foad via llvm-commits
- [llvm] 1a229a9 - X86: Preserve the dead flag clobber when expanding dynamic allocas (#227702)
via llvm-commits
- [llvm] [SLP]Fix regressions after the fmul/fadd fusion cost changes (PR #228720)
Tomas Matheson via llvm-commits
- [llvm] [VectorCombine] Remove dump() uses (NFC) (PR #229041)
Kamlesh Kumar via llvm-commits
- [polly] [Polly] Consider the preconditions of the block of a user assumption (PR #226473)
Michael Kruse via llvm-commits
- [llvm] [BranchFolding] Ignore debug instructions in singleton cond branch check (PR #228715)
Jeremy Morse via llvm-commits
- [llvm] 5fe6781 - [TableGen] Avoid crashing on a missing instruction pattern result (#228075)
via llvm-commits
- [lld] [RISCV] Disable gp relaxation if part of object unreachable (PR #72655)
Nemanja Ivanovic via llvm-commits
- [llvm] df6ef90 - CodeGen: Prefer getting the Triple from the Module (#228682)
via llvm-commits
- [llvm] PowerPC: Promote 32-bit ops to 64-bit in place for EXTSW elimination (PR #229054)
Matt Arsenault via llvm-commits
- [llvm] [GlobalISel][TableGen] Support for typed G_FCONSTANT Imm in MIR-patttern (PR #219563)
Matt Arsenault via llvm-commits
- [clang] [llvm] Revert "Revert "[AMDGPU] Disable V_WMMA_F{16|32}_16X16X128_[BF]P8_[BF]P8 on gfx1250-strict"" (PR #229044)
via llvm-commits
- [llvm] 4ab5e4e - Revert "Revert "[AMDGPU] Disable V_WMMA_F{16|32}_16X16X128_[BF]P8_[BF]P8 on gfx1250-strict"" (#229044)
via llvm-commits
- [llvm] [GlobalISel][NFC] Explicitly narrow conversions in two combine rules (PR #215210)
via llvm-commits
- [flang] [llvm] [mlir] [OpenMP] Honor barrier and last iteration flags in loop lowering (PR #224040)
Nicole Aschenbrenner via llvm-commits
- [llvm] [TableGen][NFC] Explicitly narrow ULEB128 values in generated decodeInstruction (PR #215211)
via llvm-commits
- [llvm] [Offload][OpenMP] Register RPC callback through liboffload (PR #229057)
Alex Duran via llvm-commits
- [llvm] [TableGen][NFC] Explicitly narrow size_t to unsigned in generated matcher tables (PR #215212)
via llvm-commits
- [llvm] [offload][omp] Use liboffload to initialize devices (PR #226442)
Alex Duran via llvm-commits
- [llvm] [SPIR-V] Fix and enforce minimum versions of decoration operands (PR #229058)
Arseniy Obolenskiy via llvm-commits
- [llvm] [X86] Fix Windows tail calls on 64-bit (PR #229051)
Alastair Houghton via llvm-commits
- [clang] [llvm] [IRBuilder] Drop redundant BB argument for insert point ctor (PR #229052)
Antonio Frighetto via llvm-commits
- [llvm] [AMDGPU] Restore the default rounding mode before calls and returns (PR #222247)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU][GlobalISel] Avoid 64-bit width V_BFE (PR #220653)
Arseniy Obolenskiy via llvm-commits
- [llvm] [SystemZ][z/OS] Add AMODE to PR symbols (PR #223541)
Ulrich Weigand via llvm-commits
- [llvm] [libsycl] SYCL Extension naming policy. (PR #226967)
Kseniya Tikhomirova via llvm-commits
- [llvm] [WIP][TargetLowering] Prefer 'r' over 'm' for "rm" inline asm constraints (PR #214061)
Bill Wendling via llvm-commits
- [llvm] [VPlan] Strip Idx arg in replaceUsesWithIf (NFC) (PR #229061)
Ramkumar Ramachandra via llvm-commits
- [llvm] [CodeGen] Prefer 'r' over 'm' for foldable "rm" inline asm operands (PR #214061)
Bill Wendling via llvm-commits
- [llvm] [LV] Add support for widening loads/stores to a UF x VF (PR #217670)
Benjamin Maxwell via llvm-commits
- [llvm] [NVPTX] Handle vector splats nested in aggregate initializers (PR #229073)
via llvm-commits
- [llvm] Loopidiom wchar bug (PR #229074)
Ömer Sinan Ağacan via llvm-commits
- [llvm] [SPIR-V] Lower OpenCL get_kernel_* queries to OpGetKernel* instructions (PR #227455)
Arseniy Obolenskiy via llvm-commits
- [llvm] [KnownFPClass] Fix fsub input denormal handling (PR #225605)
Matt Arsenault via llvm-commits
- [llvm] [MemDep] Stop traversal if load operand is a constant (PR #223456)
via llvm-commits
- [llvm] [AMDGPU] Skip IEEE/DX10Clamp inline compatibility check on targets without them (PR #228389)
Mirko Brkušanin via llvm-commits
- [llvm] [AMDGPU] Do not treat VOP3P `op_sel_hi` as a `dst_sel` forwarding hazard (PR #228486)
Alan Li via llvm-commits
- [llvm] [X86] Remove an unused MVT::Other result from VP2INTERSECT. (PR #228936)
Phoebe Wang via llvm-commits
- [llvm] [BasicAA] Reuse definitive cached results before object analysis (PR #228790)
Antonio Frighetto via llvm-commits
- [llvm] [AMDGPU][NFC] Prepare the MFMA hazard checks for per-producer windows (PR #225988)
Brendon Cahoon via llvm-commits
- [llvm] [Bazel] Fixes 0e7e59d (PR #229085)
via llvm-commits
- [llvm] [Transforms][Utils] Add LoopSplit for iteration-space loop splitting (PR #217232)
Florian Hahn via llvm-commits
- [llvm] 201b446 - [AMDGPU] Restore the default rounding mode before calls and returns (#222247)
via llvm-commits
- [llvm] [FastISel] Ignore debug uses when folding a load (PR #228719)
Aochang Liu via llvm-commits
- [llvm] [VectorCombine] Compose fixed shuffles through widening bitcasts (PR #229086)
Joyoungjin via llvm-commits
- [llvm] [libsycl] Fix and test sycl::item conversion (PR #227692)
Kseniya Tikhomirova via llvm-commits
- [llvm] [AMDGPU] SIPostRA16BitMovFolding optimize with forwarding mov source (PR #228576)
Guo Chen via llvm-commits
- [clang] [llvm] Add a generator for Unicode properties. (PR #229072)
via llvm-commits
- [llvm] [VPlan] Add visited set for onlyFirstLaneUsed() (PR #228438)
Florian Hahn via llvm-commits
- [llvm] [BOLT][AArch64] Port X86 --custom-allocation-vma test to AArch64 (PR #229104)
via llvm-commits
- [llvm] Improve diagnostic for malformed constant vectors (PR #216253)
via llvm-commits
- [llvm] [IR] Don't fold `bitcast` pairs through bytes into invalid casts (PR #229105)
Pedro Lobo via llvm-commits
- [llvm] [Offload] Add strong definitions of all L0 symbols for Windows (PR #228070)
Jan Trusiłło via llvm-commits
- [lldb] [llvm] [lldb] Move OS-specific UnixSignals subclasses from Plugins to Target (PR #229106)
Lexi Bromfield via llvm-commits
- [llvm] 1e9a346 - [AArch64][PeepholeOpt] Use correct register class for subreg uses from optimizeExtInstr. (#226595)
via llvm-commits
- [llvm] [AArch64][PeepholeOpt] Use correct register class for subreg uses from optimizeExtInstr. (PR #226595)
David Green via llvm-commits
- [libc] [llvm] [libc][math] Fix powf exact cases, improve its performance, and add code-size optimized float-only version. (PR #226588)
via llvm-commits
- [llvm] [AArch64] Model shifts folded into scalar logical instructions (PR #229109)
Hari Limaye via llvm-commits
- [clang] [compiler-rt] [flang] [libc] [libclc] [libcxx] [libcxxabi] [libunwind] [lld] [lldb] [llvm] [mlir] [openmp] [polly] [CMake] Bump the minimum required version to 3.31 (PR #223854)
Mehdi Amini via llvm-commits
- [llvm] [Bazel] Fixes bb878a6 (PR #229103)
Pranav Kant via llvm-commits
- [llvm] [SPIR-V][NewPM] Port SPIRVModuleAnalysis (PR #227376)
Marcos Maronas via llvm-commits
- [llvm] llvm-ml: Rectify the PROC visibility-frame token order (PR #229110)
via llvm-commits
- [llvm] [NVPTX] Implement llvm.fptrunc.round intrinsic (PR #223720)
via llvm-commits
- [llvm] [ValueTracking] Propagate KnownBits through nvvm redux and shuffle ops (PR #225424)
via llvm-commits
- [llvm] [AMDGPU] Handle register variant S_SETREG when fixing VGPR MSB clobbers (PR #218941)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU] Fold mul24 with an operand whose low 24 bits are zero (PR #224537)
LLVM Continuous Integration via llvm-commits
- [llvm] [ConstantFold] Centralize denormal-aware FP folds (PR #229111)
Lewis Crawford via llvm-commits
- [llvm] [SPIR-V] Emit ID decorations on function parameters as OpDecorateId (PR #229089)
via llvm-commits
- [llvm] [LoopIdiom] Add a range attribute to formed memset/memcpy/memmove (PR #226801)
Nathan Corbyn via llvm-commits
- [lldb] [llvm] [lldb] [bazel] Remove cycles in Bazel by merging targets (PR #229101)
Lexi Bromfield via llvm-commits
- [llvm] [NFC][LICM] Precommit mishandled per-iteration scoped alias metadata (PR #223529)
Zach Goldthorpe via llvm-commits
- [llvm] [SimplifyLibCalls] Fold more FP libcalls to intrinsics. (PR #227697)
Florian Hahn via llvm-commits
- [llvm] [X86] Improve llvm.get.active.lane.mask codegen (PR #225832)
via llvm-commits
- [llvm] [IR] Verify oracle functions are only used by llvm.speculative.load. (PR #226669)
Florian Hahn via llvm-commits
- [llvm] e0a74f4 - [AMDGPU][NFC] Prepare the MFMA hazard checks for per-producer windows (#225988)
via llvm-commits
- [llvm] [AMDGPU] Measure MFMA hazard windows at their own producers (PR #225989)
Brendon Cahoon via llvm-commits
- [lld] [lld][COFF] Don't truncate .llvm. and llvm_ section names (PR #229112)
via llvm-commits
- [llvm] [PowerPC] Don't let optimizeCompareInstr define CR0 while it is still live (PR #220999)
Daniel Moody via llvm-commits
- [llvm] [SelectionDAG] Freeze dynamic vector indices lowered through the stack (PR #225031)
Simon Pilgrim via llvm-commits
- [llvm] [GVN] Don't coerce forwarded byte values through integers (PR #226583)
Pedro Lobo via llvm-commits
- [llvm] [polly] [IRBuilder] Directly pass insertion point to ctor in more cases (PR #229118)
Nikita Popov via llvm-commits
- [llvm] [BOLT] Gadget scanner: prevent false positives due to jump tables (PR #138884)
Anatoly Trosinenko via llvm-commits
- [llvm] [mlir] [MLGO] Do not hardcode 33 eviction feature columns (PR #224598)
Mircea Trofin via llvm-commits
- [llvm] [ConstantFolding] Fold `bitinsert` and `bitextract` (PR #229115)
Pedro Lobo via llvm-commits
- [llvm] [MemDep] Stop traversal if invariant.group load operand is a constant (PR #223456)
via llvm-commits
- [clang] [llvm] [AMDGPU] Implement AsyncMark Stages (PR #220442)
Ryan Mitchell via llvm-commits
- [llvm] [SCEV] Prove min/max Pred RHS and LHS Pred min/max via decomposition. (PR #208546)
Florian Hahn via llvm-commits
- [llvm] [AMDGPU] Turn tablegen tables with single bool into filtered lists. NFC (PR #227746)
Mirko Brkušanin via llvm-commits
- [llvm] [AMDGPU][GlobalISel] Fold zext(carry) into add/sub-with-carry (PR #224093)
Mirko Brkušanin via llvm-commits
- [llvm] [LV] Don't tail-fold the epilogue with EVL-based tail-folding (PR #228452)
Hassnaa Hamdi via llvm-commits
- [llvm] [IRBuilder] Deprecate overloads with redundant BB argument (PR #229102)
Nikita Popov via llvm-commits
- [llvm] [GlobalOpt] Skip cold call analysis when no function can use coldcc (PR #227944)
Malavika Samak via llvm-commits
- [llvm] [Offload][OpenMP] Register RPC callback through liboffload (PR #229149)
Alex Duran via llvm-commits
- [clang] [llvm] [RISCV][clang] Support XSfmm ABI attributes (PR #206260)
Craig Topper via llvm-commits
- [llvm] [AArch64] Lower fixed find_last_active to LASTP with SVE2.2 (PR #222534)
Vimal Patel via llvm-commits
- [llvm] 1dafb43 - [NFC][HIP][SPIRV] Rework `-spirv-preserve-auxdata` handling. (#227875)
via llvm-commits
- [clang] [llvm] [NFC][HIP][SPIRV] Rework `-spirv-preserve-auxdata` handling. (PR #227875)
Alex Voicu via llvm-commits
- [llvm] [libsycl] Disable werror for deprecated declaration in test (PR #229145)
Nick Sarnie via llvm-commits
- [llvm] [PassBuilder] Add extra ModuleInliner/SROA/SimplifyCFG/LICM to late FullLTO passes (PR #197678)
Momchil Velikov via llvm-commits
- [llvm] [LoopUnroll] Disable using PrepareForLTO for costing (PR #229150)
Aiden Grossman via llvm-commits
- [llvm] [Bazel] Fixes 38d295d (PR #229142)
Anshul Nigham via llvm-commits
- [llvm] df91acf - [NFC][LICM] Precommit mishandled per-iteration scoped alias metadata (#223529)
via llvm-commits
- [llvm] 1b46cad - [LICM] Drop per-iteration AA tags (#223530)
via llvm-commits
- [llvm] [LICM] Drop *only* per-iteration AA tags (PR #225203)
Zach Goldthorpe via llvm-commits
- [llvm] [AMDGPU][GISel] Honor the nomerge attribute on calls (PR #229152)
Joseph Huber via llvm-commits
- [llvm] [LoopUnroll] Unroll when requested despite inlining candidates (PR #219077)
Aiden Grossman via llvm-commits
- [llvm] [AMDGPU] Add target features for artificial LDS size limit (PR #229153)
Mariusz Sikora via llvm-commits
- [llvm] [SPIR-V] Mark elementwise intrinsics IntrTriviallyScalarizable (PR #227286)
Farzon Lotfi via llvm-commits
- [llvm] [RISCV] Add mask version of vunzipe/o.v (PR #228597)
Min-Yih Hsu via llvm-commits
- [clang] [llvm] [HLSL] Implement isfinite() intrinsic (PR #225947)
Farzon Lotfi via llvm-commits
- [lld] [ELF] Error on non-TLS definition with TLS reference (PR #228658)
Fangrui Song via llvm-commits
- [llvm] [mlir] [MLGO] Make regalloc eviction input tensor shapes runtime variables (PR #224598)
Arseniy Obolenskiy via llvm-commits
- [llvm] [StandardInstrumentations] Add ExtendedIRContext and trait registration for custom IR types (PR #153171)
Arthur Eubanks via llvm-commits
- [flang] [llvm] [mlir] [flang][mlir][OpenMP] Report Fortran names for privatized target maps (PR #228195)
Jason Van Beusekom via llvm-commits
- [llvm] SystemZ: Remove stale CC live range when converting AND to RISBG (PR #229062)
Matt Arsenault via llvm-commits
- [llvm] [ValueTracking][NFC] Add haveCommonLowBits helper (PR #228168)
Sayan Sivakumaran via llvm-commits
- [llvm] [obj2yaml] Add Obj2Yaml ability to dump GOFF ESD records (PR #216151)
Yusra Syeda via llvm-commits
- [llvm] [SLSR] Skipping rewriting based on liveness (PR #218470)
Yoonseo Choi via llvm-commits
- [llvm] [X86] Promote approximate f16 vector FMA to f32 (PR #228900)
Evgenii Kudriashov via llvm-commits
- [llvm] [RISCV] Use vunzip instead of shift and truncate in the presence of Zvzip (PR #228298)
Min-Yih Hsu via llvm-commits
- [llvm] [SLP] Advance insertion point over vectorized nodes. (PR #227835)
via llvm-commits
- [llvm] [RISCV][Disassembler] Reject GPR pairs with x16-x31 under RVE (PR #229166)
Sam Elliott via llvm-commits
- [llvm] [SelectionDAG] Add opcode class metadata to DOT nodes (PR #228965)
Matt Arsenault via llvm-commits
- [llvm] [AMDGPU] Mark constant HW-reg reads and MBCNT cheap-as-move to avoid nonlocal CSE (PR #228500)
Nick Riasanovsky via llvm-commits
- [llvm] [amdgpu] Add HasBufferInv AMDGPUSubtargetFeature (PR #226254)
Matt Arsenault via llvm-commits
- [clang] [compiler-rt] [llvm] [AArch64] Extend FMV support to more features (PR #229139)
Alexandros Lamprineas via llvm-commits
- [llvm] [RISCV][Support] Fix instruction-cache invalidation in Memory::InvalidateInstructionCache (PR #216059)
Sam Elliott via llvm-commits
- [llvm] SystemZ: Don't convert AND to RISBG if CC is live (PR #229063)
Matt Arsenault via llvm-commits
- [lld] [lld][MachO] Parallelize parsing relocations for input files. (PR #223795)
Liza Burakova via llvm-commits
- [llvm] [AArch64] Preserve store width for ptr32 address spaces (PR #229181)
Daniel Paoliello via llvm-commits
- [llvm] [flang-rt][test] Modify requirements for Driver/safe-trampoline-gnustack.f90 (PR #227851)
Tulio Magno Quites Machado Filho via llvm-commits
- [llvm] [llvm-profgen] Parse hex address suffixes in perf scripts (PR #219268)
Rahman Lavaee via llvm-commits
- [llvm] DAG: Gracefully diagnose missing floating-point state libcalls (PR #228426)
Serge Pavlov via llvm-commits
- [llvm] [AMDGPU] Add `hasSelectableWaveSize` to TargetParser (PR #229160)
Krzysztof Drewniak via llvm-commits
- [llvm] ad3539c - [amdgpu] Add HasBufferInv AMDGPUSubtargetFeature (#226254)
via llvm-commits
- [llvm] d7b96ba - [RISCV] Use vunzip instead of shift and truncate in the presence of Zvzip (#228298)
via llvm-commits
- [llvm] [AMDGPU] Update LiveVariables in SILowerControlFlow::combineMasks (PR #217844)
Matt Arsenault via llvm-commits
- [llvm] [libc][bazel] Add missing MPFR test targets for fmax/fmin family (PR #223867)
Usef Ahmed via llvm-commits
- [llvm] [AMDGPU][NFC] Clean up packed convert tests (PR #229178)
Chinmay Deshpande via llvm-commits
- [llvm] [OpenMP] Fix debug locations in GPU reductions (PR #228622)
Amadeu Moya via llvm-commits
- [llvm] [Mips][nanoMIPS] Add initial nanoMIPS target support (PR #223032)
via llvm-commits
- [clang] [llvm] [mlir] [AMDGPU] Introduce builtin, intrinsic and ROCDL instruction for buffer_inv (PR #228497)
Dan Zimmerman via llvm-commits
- [llvm] [hlsl][spirv] Add late pass in spirv backend to remove dead resources (PR #225270)
Helena Kotas via llvm-commits
- [llvm] [openmp] [OpenMP][offload] Fix single-thread SPMD cross-team reductions (PR #228216)
Robert Imschweiler via llvm-commits
- [llvm] 639e8db - [ConstantFold] Don't let noipa block global address equality folding (#228186)
via llvm-commits
- [llvm] [ConstantFold] Don't let noipa block global address equality folding (PR #228186)
David Blaikie via llvm-commits
- [llvm] [msan][test] Add tests for llvm.vector.partial.reduce.add/fadd (PR #229196)
Thurston Dang via llvm-commits
- [llvm] c29426f - [RISCV][Disassembler] Reject GPR pairs with x16-x31 under RVE (#229166)
via llvm-commits
- [llvm] [mlir] [mlir][OpenMP] Allow multi-block omp.iterator regions (PR #229198)
via llvm-commits
- [llvm] 08b3dd3 - SystemZ: Don't convert AND to RISBG if CC is live (#229063)
via llvm-commits
- [llvm] [VPlan] Encode single-scalar-ness of VPInstruction in a field (NFC). (PR #227761)
Ramkumar Ramachandra via llvm-commits
- [llvm] MIR: Serialize MachineBasicBlock::IsCleanupFuncletEntry (PR #227194)
Sam Elliott via llvm-commits
- [llvm] a3ebd6a - [SPIR-V] Lower get_fence builtin to OpGenericPtrMemSemantics (#229023)
via llvm-commits
- [llvm] 2e52adb - [SLP][AMDGPU][NFC] Precommit test for alternate node fmul cost (#229185)
via llvm-commits
- [llvm] [SLP][AMDGPU][NFC] Precommit test for alternate node fmul cost (PR #229185)
via llvm-commits
- [llvm] MachineLICM: Stop checking kill flags in register pressure estimate (PR #229201)
Matt Arsenault via llvm-commits
- [flang] [llvm] [flang][cuda] share runtime type info between host and device under managed memory (PR #229213)
via llvm-commits
- [llvm] [SPIR-V] Lower clang __to_global, __to_local and __to_private builtins (PR #229211)
via llvm-commits
- [llvm] [VPlan] Avoid recursive-fan-out of usesFirstLaneOnly (PR #229130)
Ramkumar Ramachandra via llvm-commits
- [llvm] [CI] Exclude CIR from Windows premerge testing (PR #229208)
Aiden Grossman via llvm-commits
- [llvm] [libc][bazel] Bazel fixes for 57732ac15d3e2305803c45ca748e4e1443bdcddc (PR #229210)
via llvm-commits
- [llvm] [SLP] Partially revert #224931 for alternate nodes (PR #228659)
Alexey Bataev via llvm-commits
- [llvm] [SeparateConstOffsetFromGEP] Simplify zero operands when rebuilding indices (PR #229209)
Alex MacLean via llvm-commits
- [llvm] RuntimeLibcalls: Add function signatures to RTLIB. (PR #226264)
Kiroo via llvm-commits
- [llvm] [Bazel] Fix MLIR bazel build (PR #229217)
Anshul Nigham via llvm-commits
- [llvm] llc: Verify MIR outputs by default (PR #228248)
Matt Arsenault via llvm-commits
- [llvm] [X86] Implement preferVectorizedNonPowerOfTwoTypeBreakdown() for X86 (PR #227169)
Ryan Buchner via llvm-commits
- [llvm] [AMDGPU] Reserve ENABLE_WAVEFRONT_SIZE32 on gfx125x (PR #229162)
Chinmay Deshpande via llvm-commits
- [llvm] [debuginfo] Add DW_AT_type support to DW_TAG_string_type DIEs (1 of 3) (PR #217419)
via llvm-commits
- [llvm] 1def37a - DAG: Fix error message casing to match style policy (#229212)
via llvm-commits
- [llvm] DAG: Fix error message casing to match style policy (PR #229212)
Matt Arsenault via llvm-commits
- [llvm] [CodeGen] Eliminate redundant spills at control-flow joins (PR #226852)
Yaxun Liu via llvm-commits
- [llvm] [X86] Add IR pass to insert sfence after non-temporal stores (PR #229240)
James Y Knight via llvm-commits
- [llvm] [GlobalISel] Add G_INSERT_VECTOR_ELT to computeNumSignBits (PR #213480)
Tim Gymnich via llvm-commits
- [llvm] 6f29c16 - [DWARFLinker] Add missing DebugInfoDWARFLowLevel dependency (#229222)
via llvm-commits
- [llvm] [DWARFLinker] Add missing DebugInfoDWARFLowLevel dependency (PR #229222)
Adrian Prantl via llvm-commits
- [llvm] [SPIR-V] Split wide vector intrinsics during legalization (PR #227287)
Joshua Batista via llvm-commits
- [compiler-rt] [NFC][TSan] Use in-class member initializers in tsan_report.h (PR #228780)
Andrew Lazarev via llvm-commits
- [llvm] [AArch64][GlobalISel] Improve MOVI constant splat lower (PR #229241)
David Green via llvm-commits
- [llvm] [DebugInfo] Add DILayerLoc/DILayerLocList and DILocation irlayers operand (PR #215666)
Alexander Yermolovich via llvm-commits
- [compiler-rt] [llvm] [libFuzzer] Stop fuzzing if no new corpus were found within a specified period (PR #176177)
via llvm-commits
- [llvm] [X86] Add llvm.x86.movnt, an opaque non-temporal store intrinsic (PR #228875)
James Y Knight via llvm-commits
- [llvm] [NVPTX][TTI] Update NVPTX TTI to enable f32x2 vectorization (PR #222994)
Daniel Donenfeld via llvm-commits
- [clang] [llvm] [HLSL] Add HLSL unpack intrinsics (PR #228518)
via llvm-commits
- [compiler-rt] [llvm] Reland "[compiler-rt][asan] Add asan checks for __builtin_assume_dereferencable" (PR #224163)
Vitaly Buka via llvm-commits
- [clang] [llvm] [clang][Inliner] Add opt-in inliner hints and implementation-file functions (PR #228074)
Kuba Mracek via llvm-commits
- [llvm] Don't constant-fold equality comparisons on global declarations. (PR #229243)
Eli Friedman via llvm-commits
- [llvm] [MachineScheduler] Fix stale read-undef flags when rescheduling physreg copies (PR #227897)
Nick Riasanovsky via llvm-commits
- [llvm] [SPARC] Prevent RESTORE from consuming an indirect call's target (PR #228198)
Matt Arsenault via llvm-commits
- [llvm] [AMDGPU] Track va_ssrc / va_sdst hazards (PR #225810)
Matt Arsenault via llvm-commits
- [llvm] [SLP] Don't rely on VectorizableTree in canBuildSplitNode() when called from tryToReduce() (PR #220014)
Alexey Bataev via llvm-commits
- [compiler-rt] [sanitizer] Clear pending dlerror on thread exit before unregistering (PR #228930)
LLVM Continuous Integration via llvm-commits
- [llvm] [AMDGPU] Split the true16 fcopysign pattern by uniformity (PR #226898)
Changpeng Fang via llvm-commits
- [llvm] 8374529 - [AMDGPU] Split the true16 fcopysign pattern by uniformity (#226898)
via llvm-commits
- [llvm] [MachineOutliner] Attribute outlined calls to the candidate's last call (PR #229260)
Kyungwoo Lee via llvm-commits
- [compiler-rt] [NFC][lsan] Disable HWASan on key_destructor in detached_thread_dlsym.c (PR #229266)
Vitaly Buka via llvm-commits
- [llvm] [SLP][X86]Vectorize alternate fadd/fsub fused with fmul (fmaddsub/fmsubadd) (PR #229268)
Alexey Bataev via llvm-commits
- [llvm] A few improvement in fcmla pattern recognitions (PR #173818)
Yichao Yu via llvm-commits
- [compiler-rt] 179a0cc - [NFC][lsan] Disable HWASan on key_destructor in detached_thread_dlsym.c (#229266)
via llvm-commits
- [llvm] [InstCombine] Fold sext of an nsw binop on trunc nsw into a wide nsw … (PR #228530)
Alexey Bataev via llvm-commits
- [clang] [llvm] [Clang][AMDGPU] Add nontemporal scoped atomic load/store and merge relaxed atomics (PR #229207)
Joseph Huber via llvm-commits
- [libc] [llvm] [libc] Rewrite libc GPU code to use rpc::dispatch where possible (PR #181727)
m][sko via llvm-commits
- [llvm] [InstCombine] Fix ProfCheck for canonicalizeClampLike (PR #229269)
Aiden Grossman via llvm-commits
- [llvm] [SelectionDAG] Implement SDTCisPtrTy in SDNodeInfo::verifyNode. (PR #229270)
Craig Topper via llvm-commits
- [llvm] [CodeGen] Ignore debug instructions in register scavenging (PR #229271)
Yongqiang Tian via llvm-commits
- [llvm] ad39cc1 - [InstCombine] Fix ProfCheck for canonicalizeClampLike (#229269)
via llvm-commits
- [flang] [llvm] [flang][cuda][semantics] Reject unguarded host-only and device-only calls in CUDA HOST,DEVICE procedures (PR #228176)
Eugene Epshteyn via llvm-commits
- [llvm] [SPIRV] Default OpenCL 2.0 atomic RMW builtins to Device scope (PR #229285)
Alexander Shaposhnikov via llvm-commits
- [llvm] [OpenMP][Offload] Fix binary search in MemoryManager.h (PR #229290)
Muyao Xiao via llvm-commits
- [llvm] [VPlan] Implement VPlan-based unit-strideness speculation (PR #182595)
Andrei Elovikov via llvm-commits
- [llvm] [NVPTX] Add Custom ABI (abi_preserve) support via function attributes (PR #227499)
Akshay Deodhar via llvm-commits
- [llvm] [AMDGPU] Rejoin terminating branches that call noreturn functions (PR #229291)
Joseph Huber via llvm-commits
- [llvm] [X86][AMX] Fix use iteration when reloading tiles in FastPreTileConfig (PR #228724)
Aochang Liu via llvm-commits
- [flang] [llvm] [Opened by mistake] (PR #171112)
Muyao Xiao via llvm-commits
- [llvm] [AtomicExpand] Add elementwise expansion (PR #196464)
Yonah Goldberg via llvm-commits
- [clang] [llvm] [TargetParser][Clang] Parse Intel GPU architecture names (PR #229247)
Srividya Sundaram via llvm-commits
- [clang] [llvm] [Clang][SYCL] Infer the SPIR-V target from an Intel offload architecture (PR #229248)
Srividya Sundaram via llvm-commits
- [clang] [llvm] [Clang][SYCL] Reject an unknown offload architecture on a SPIR-V target (PR #229249)
Srividya Sundaram via llvm-commits
- [llvm] [Draft] Data flow based shrink wrapping (PR #191784)
Piyou Chen via llvm-commits
- [clang] [llvm] [LLVMABI][AARCH64] Added aggregate return support (PR #229202)
via llvm-commits
- [llvm] [AArch64][GlobalISel] Don't let debug instructions change register classes (PR #228078)
Aochang Liu via llvm-commits
- [llvm] [Verifier] Don't treat noipa callees as non-inlinable in the !dbg check (PR #228241)
LLVM Continuous Integration via llvm-commits
- [compiler-rt] 197fb4b - [NFC][TSan] Use in-class member initializers in tsan_report.h (#228780)
via llvm-commits
- [llvm] [MachineOutliner] Preserve debug locations on outlined call sequences (PR #224189)
Ian Anderson via llvm-commits
- [llvm] [GlobalISel] Salvage dead integer constants in debug values (PR #229297)
Yongqiang Tian via llvm-commits
- [llvm] [DAG][X86] Bitfield insertion can combine into SHL + SHRD (PR #220182)
Pratyay Pande via llvm-commits
- [compiler-rt] f42afb1 - [NFC][TSan] Move ObtainCurrentStack out of ThreadRegistryLock scope (#228794)
via llvm-commits
- [compiler-rt] [NFC][TSan] Move ObtainCurrentStack out of ThreadRegistryLock scope (PR #228794)
Vitaly Buka via llvm-commits
- [clang] [llvm] [mlir] [CMake] Make find_package(Clang) load MLIR when ClangIR uses a full MLIR (PR #228366)
via llvm-commits
- [clang] [llvm] [CMake] Link MLIR if CLANG_ENABLE_CIR is ON (PR #226763)
via llvm-commits
- [llvm] [InstrProf] Inline ir-basic.proftext into test (PR #219329)
Kyungwoo Lee via llvm-commits
- [llvm] Introduce check-first vectorization for early-exit loops. (PR #227201)
Arjun H Kumar via llvm-commits
- [llvm] [AArch64] Declare command line options in TableGen (PR #229303)
Fangrui Song via llvm-commits
- [llvm] [GlobalISel] Ignore debug instructions in extracted-load scan budget (PR #229304)
Yongqiang Tian via llvm-commits
- [llvm] PR 213007 fixup (PR #228980)
Sushant Gokhale via llvm-commits
- [llvm] [IR] Check for integer overflow in ShuffleVectorInst::isValidOperands() (PR #229306)
Thurston Dang via llvm-commits
- [llvm] [KnownFPClass] Correct denormal mode handling for `KnownFPClass::minMaxLike` (PR #220148)
Eric Ross via llvm-commits
- [llvm] [RISCV] Prefix option names with riscv-, turn disable- to positive options (PR #228922)
Fangrui Song via llvm-commits
- [compiler-rt] [TSan] Lock ScopedErrorReportLock before slot and thread_registry locks (PR #228614)
Vitaly Buka via llvm-commits
- [llvm] [X86] Require contract on FMUL while folding FMADDSUB to VFMULC (PR #229309)
Rohan Shenoy via llvm-commits
- [llvm] [flang-rt] Fix BACKSPACE crash on files without trailing newline (PR #227661)
Kaviya Rajendiran via llvm-commits
- [compiler-rt] 78c158c - [TSan] Lock ScopedErrorReportLock before slot and thread_registry locks (#228614)
via llvm-commits
- [llvm] [DAGCombiner] Avoid illegal scalar nodes in estimate refinement (PR #219394)
via llvm-commits
- [llvm] [AMDGPU] Lower loads and stores for address space 13 (PR #209541)
Gheorghe-Teodor Bercea via llvm-commits
- [llvm] [KnownFPClass] Improve known class deductions for exp/exp2/exp10 (PR #217526)
Eric Ross via llvm-commits
- [llvm] [CodeGenPrepare] Reject address combining if a new node uses a value that does not dominate it (PR #226968)
Aochang Liu via llvm-commits
- [llvm] [NFC][Target] Simplify checking target arch (PR #229312)
Ben Wu via llvm-commits
- [compiler-rt] [llvm] [msan] Add interceptors for fortified memory functions (__memcpy_chk, etc.) (PR #206702)
Austin Schuh via llvm-commits
- [llvm] [TableGen] Generate partial mappings (PR #229314)
via llvm-commits
- [compiler-rt] [NFC][TSan] Allocate ScopedReport as a stack variable (PR #228637)
Vitaly Buka via llvm-commits
- [compiler-rt] 695d0e1 - [NFC][TSan] Allocate ScopedReport as a stack variable (#228637)
via llvm-commits
- [llvm] 0e5770c - [KnownFPClass] Improve known class deductions for exp/exp2/exp10 (#217526)
via llvm-commits
- [compiler-rt] [TSan] Defer symbolization in AddStack and AddSleep to SymbolizeStackElems (PR #228795)
Vitaly Buka via llvm-commits
- [llvm] [IR] Define empty and malformed !callees semantics (PR #221550)
Mian Miftah via llvm-commits
- [llvm] [KnownFPClass] Fix KnownFPClass::frexp subnormal mode handling (PR #219114)
Eric Ross via llvm-commits
- [llvm] 0837320 - [KnownFPClass] Fix KnownFPClass::frexp subnormal mode handling (#219114)
via llvm-commits
- [llvm] [RISCV] Simplify float-point boundary comparisons with FCLASS (PR #228984)
Craig Topper via llvm-commits
- [compiler-rt] 496a424 - [TSan] Defer symbolization in AddStack and AddSleep to SymbolizeStackElems (#228795)
via llvm-commits
- [llvm] [AMDGPU] Ignore debug uses when folding REG_SEQUENCE copies (PR #229299)
Aochang Liu via llvm-commits
- [compiler-rt] [NFC][TSan] Check IsFiredSuppression before ScopedReport in ReportErrnoSpoiling (PR #228811)
Vitaly Buka via llvm-commits
- [llvm] [CodeGen] Let RegAllocFast lower tied operands by default (PR #229310)
Alexis Engelke via llvm-commits
- [lld] [lld][WebAssembly] Error on TLS relocations against shared library symbols (PR #229308)
Sam Clegg via llvm-commits
- [clang] [llvm] [AMDGPU] Fix out-of-bounds read of a trailing '%' in printf formats (PR #229088)
Steffen Larsen via llvm-commits
- [llvm] AArch64/GlobalISel: Drop redundant IMPLICIT_DEFs before the ptrauth pseudos (PR #227217)
Cullen Rhodes via llvm-commits
- [llvm] [AMDGPU] Move call site info when SI_CALL_ISEL is replaced (PR #229318)
Mian Miftah via llvm-commits
- [llvm] [AMDGPU] Preserve call info when lowering SI_CALL_ISEL (PR #192200)
Mian Miftah via llvm-commits
- [clang] [llvm] [clang][OpenMP] Add no-loop SPMD kernel promotion (PR #224041)
Nicole Aschenbrenner via llvm-commits
- [compiler-rt] [NFC][TSan] Remove unnecessary ThreadRegistryLock in signal and mutex-held reports (PR #228812)
Vitaly Buka via llvm-commits
- [llvm] AArch64: Error on unsupported calls with no arguments (PR #229319)
Matt Arsenault via llvm-commits
- [llvm] [MIR] Resolve callSites and calledGlobals blocks by number (PR #229320)
Mian Miftah via llvm-commits
- [llvm] [GlobalISel] Support fixed vectors in constant-fold combines. (PR #220876)
Vikash Gupta via llvm-commits
- [llvm] [RISCV][GISel] Add CFProtectionBranch support to legalizeBRJT. (PR #229311)
Matt Arsenault via llvm-commits
- [compiler-rt] c428394 - [NFC][TSan] Remove unnecessary ThreadRegistryLock in signal and mutex-held reports (#228812)
via llvm-commits
- [llvm] [SPIRV] Fix invalid SPIR-V for loops with switch headers (PR #229289)
Arseniy Obolenskiy via llvm-commits
- [llvm] [BOLT] Move target-specific relocation logic into target handlers (PR #217926)
via llvm-commits
- [llvm] [AMDGPU] Return false, instead of asserting, when merging mov_b16s (PR #229265)
Jay Foad via llvm-commits
- [llvm] [AMDGPU] Fix amdgcn.mbcnt known bits conflicting with the range attribute (PR #227725)
Teja Alaghari via llvm-commits
- [llvm] [NVPTX] Copy 16 byte aligned small memcpy with v4 loads and stores (PR #229326)
via llvm-commits
- [llvm] [Analysis] Remove unused AssumptionCacheTracker::lookupAssumptionCache (NFC) (PR #229327)
Kazu Hirata via llvm-commits
- [llvm] [ARM][GISel] Enable extended LLT (PR #191830)
David Green via llvm-commits
- [llvm] [CodeGen][RISCV][AArch64] Added signed extension support to TypePromotion (PR #226304)
Sam Parker via llvm-commits
- [llvm] [SelectionDAG] Attach call info to the call, not to a copy before it (PR #229329)
Mian Miftah via llvm-commits
- [llvm] [SLP] Cost reduce.add(mul(ext, ext)) as a dot-product reduction (PR #224066)
Madhur Amilkanthwar via llvm-commits
- [clang] [llvm] [mlir] [AMDGPU] Fix missing cache bypass scope on seq_cst buffer fat pointer atomics (PR #213585)
Arseniy Obolenskiy via llvm-commits
- [llvm] [SPIRV] Handle loads from indirect args in the presence of `SPV_KHR_untyped_pointers` (PR #229184)
Alex Voicu via llvm-commits
- [llvm] [CodeGen] Clear call site info when resetting a MachineFunction (PR #229331)
Mian Miftah via llvm-commits
- [llvm] [SDAG] Allow S/Zext in select-abd pattern (PR #191207)
via llvm-commits
- [llvm] [CodeGen] Ignore debug instructions in EarlyIfConversion scan budget (PR #224495)
Florian Hahn via llvm-commits
- [llvm] [AMDGPU] Make globally-addressable-scratch opt-in (PR #189555)
Sameer Sahasrabuddhe via llvm-commits
- [llvm] [AMDGPU] Skip functions with convergence tokens in AtomicOptimizer (PR #227929)
Sameer Sahasrabuddhe via llvm-commits
- [llvm] 6973db5 - [SelectionDAG] Handle constants zero-folding in SimplifyMultipleUseDemandedBits (#224882)
via llvm-commits
- [llvm] [SelectionDAG] Handle constants zero-folding in SimplifyMultipleUseDemandedBits (PR #224882)
Harrison Hao via llvm-commits
- [llvm] Attributor: Use NullPointerIsDefined for null nocapture check (PR #229333)
Matt Arsenault via llvm-commits
- [llvm] 63239df - [NFC][Target] Simplify checking target arch (#229312)
via llvm-commits
- [flang] [llvm] [Flang][Runtime] Add fast path for formatted real input (PR #229335)
via llvm-commits
- [llvm] 93b307b - [Analysis] Remove unused AssumptionCacheTracker::lookupAssumptionCache (NFC) (#229327)
via llvm-commits
- [llvm] X86: Mark EFLAGS clobbers dead in global base register setup (PR #229337)
Matt Arsenault via llvm-commits
- [llvm] feat: add flag to disable known bits analysis (PR #228000)
Luisa Cicolini via llvm-commits
- [llvm] AMDGPU: Remove dead check for amdgpu-unsafe-fp-atomics attribute (PR #229338)
Matt Arsenault via llvm-commits
- [llvm] [LV] Only fold outer extend into narrow mul if it cannot wrap. (PR #229230)
Benjamin Maxwell via llvm-commits
- [llvm] [GlobalISel] Initialize MFAM in RegBankSelect RepairingPlacement (PR #229050)
Matt Arsenault via llvm-commits
- [llvm] [CodeGen] Expand unsupported interleave to shuffles (PR #210494)
Kamlesh Kumar via llvm-commits
- [llvm] e5893c1 - [AMDGPU] Fix amdgcn.mbcnt known bits conflicting with the range attribute (#227725)
via llvm-commits
- [llvm] [GVN] Use willNotFreeBetween in loop-load PRE (PR #228033)
Antonio Frighetto via llvm-commits
- [llvm] b905757 - [SPIRV] Handle loads from indirect args in the presence of `SPV_KHR_untyped_pointers` (#229184)
via llvm-commits
- [llvm] [llvm-objcopy] Route section copies through writeSectionContents (PR #217974)
James Henderson via llvm-commits
- [clang] [llvm] [AArch64] Add support for C2 CPUs (PR #229055)
via llvm-commits
- [llvm] [Offload] Add GenericProfilerTy abstraction and APITypes extensions (PR #225639)
Jan Patrick Lehr via llvm-commits
- [polly] [Polly] Keep the detection context of the expanded region (PR #229341)
Timur Baidusenov via llvm-commits
- [llvm] Add local optimization pass for compiler certification (PR #229342)
Arul Samuel Inbaraj via llvm-commits
- [llvm] [CodeGen] Recompute intervals after erasing subregister IMPLICIT_DEF (PR #218348)
Dominik Montada via llvm-commits
- [llvm] [ConstraintElim] Forget loop SCEV when folding an exiting condition. (PR #213945)
Florian Hahn via llvm-commits
- [llvm] [GlobalISel] Migrate remaining generic wip_match_opcode combines to MIR-pattern. (PR #222851)
Vikash Gupta via llvm-commits
- [llvm] [CodeGen][RegCoalescer] Shrink non-PHI live-through subranges (PR #216297)
Dominik Montada via llvm-commits
- [compiler-rt] [compiler-rt][ubsan] Link -ldl in hsa-missing.cpp (PR #229343)
Riyaz Ahmad via llvm-commits
- [compiler-rt] [compiler-rt][ubsan] Add powerpc64le support to offload UBSan HSA interceptors (PR #228007)
Riyaz Ahmad via llvm-commits
- [llvm] [Bazel] Fixes fc90bd3 (PR #229344)
via llvm-commits
- [polly] [Polly] Consider all conditions in loop for its unbounded parameter values (PR #226549)
Timur Baidusenov via llvm-commits
- [llvm] [LV] Drop prefetches from vector loops (PR #226895)
Ramkumar Ramachandra via llvm-commits
- [llvm] [GlobalISel] Ensure integer trunc and anyext in call lowering. (PR #229346)
David Green via llvm-commits
- [llvm] [AArch64][SPIRV][GlobalISel] Migrate target-specific wip_match_opcode combines to MIR-pattern (PR #222903)
Vikash Gupta via llvm-commits
- [llvm] ff4c97d - AMDGPU: Remove dead check for amdgpu-unsafe-fp-atomics attribute (#229338)
via llvm-commits
- [libc] [llvm] [libc] Moves __cxa_thread_atexit_impl into proper entrypoint (PR #229347)
via llvm-commits
- [llvm] 3535511 - [SPIRV] Default OpenCL 2.0 atomic RMW builtins to Device scope (#229285)
via llvm-commits
- [llvm] 97c2b12 - [SPIRV] Fix invalid SPIR-V for loops with switch headers (#229289)
via llvm-commits
- [llvm] [AArch64][NeoverseV1] Fix integer divide throughput (PR #224592)
Julien Villette via llvm-commits
- [llvm] [LV] Recognize safe loads through identical GEPs (PR #228797)
Ramkumar Ramachandra via llvm-commits
- [llvm] Add local optimization pass (PR #229348)
Arul Samuel Inbaraj via llvm-commits
- [llvm] [ConstantTime][LLVM] Add llvm.ct.select intrinsic with generic SelectionDAG lowering (PR #166702)
Thomas Bizet via llvm-commits
- [llvm] [CodeGen] Drop tail-local IR values from memoperands in tail duplication (PR #226768)
MMS IT GmbH via llvm-commits
- [llvm] [AArch64] Reorder callee-saved spills to maximise load/store register groupings (PR #229349)
Kieran B via llvm-commits
- [llvm] [RegAlloc] Consider rematerialization over CSR use (PR #206756)
Lukas Sommer via llvm-commits
- [llvm] [VPlan] Do not reuse scalar NSW on reconstructed widened induction mul/add (PR #229350)
via llvm-commits
- [llvm] [llubi] Implements arithmetic and reduction VP intrinsics (PR #229351)
Zhige Chen via llvm-commits
- [llvm] ef18054 - [LAA] Avoid duplicate runtime-check group members (#226829)
via llvm-commits
- [llvm] [llubi] Initial support for vector predication intrinsics (PR #212093)
Zhige Chen via llvm-commits
- [clang] [llvm] AMDGPU: Improve optimization remarks for atomic lowering (PR #229352)
Matt Arsenault via llvm-commits
- [llvm] [X86] Remove x86-widen-shift flag (PR #229353)
Simon Pilgrim via llvm-commits
- [llvm] [Mips][BranchFolding] Enable safe tail merging (PR #225382)
Jiaxun Yang via llvm-commits
- [llvm] [TableGen][AMDGPU][AsmParser] Catch and fix ambiguous instructions (PR #211093)
Ivan Kosarev via llvm-commits
- [llvm] 08652b2 - [AArch64][LSR] Prefer pointer IVs for SVE accesses that need splitting (#228510)
via llvm-commits
- [llvm] [TableGen] Support MachineInstr rewriting in CompressInstEmitter (PR #224212)
Jiaxun Yang via llvm-commits
- [compiler-rt] [sanitizer_common] Add tests for the C11 <threads.h> functions (PR #229066)
Krzysztof Jędruczyk via llvm-commits
- [llvm] [VPlan] Add getLastOperand helper (NFC). (PR #229354)
Florian Hahn via llvm-commits
- [llvm] [AMDLIBM] Map new vector calls in VecFuncs (PR #224614)
Rohit Aggarwal via llvm-commits
- [llvm] 92d7d31 - [AArch64][SPIRV][GlobalISel] Migrate target-specific wip_match_opcode combines to MIR-pattern (#222903)
via llvm-commits
- [compiler-rt] [compiler-rt] Build the builtins library for AArch64 big endian (PR #225996)
Peter Smith via llvm-commits
- [clang] [llvm] [InstCombine] Combine ((ext(A) * ext(B)) >> N) -> ext(mulh(A, B)). (PR #224311)
Ricardo Jesus via llvm-commits
- [llvm] 94eb65e - [CMake] Make find_package(Clang) load MLIR when ClangIR uses a full MLIR (#228366)
via llvm-commits
- [llvm] [AMDGPU] Add fmul.legacy select fold tests. NFC (PR #229356)
Harrison Hao via llvm-commits
- [libcxx] [llvm] [libcxx][CMake] Rework LLVM_ENABLE_ASSERTIONS interaction w. libcxx hardening (PR #229359)
Mészáros Gergely via llvm-commits
- [llvm] [NewPM] Skip result cleanup when no analyses were invalidated (NFC) (PR #229360)
Florian Hahn via llvm-commits
- [clang] [llvm] [ARM][AArch64] Introduce Armv9.8-A architecture version (PR #229361)
Jonathan Thackray via llvm-commits
- [llvm] [AArch64] Use SVE for interleaved accesses in streaming mode (PR #229366)
Kamlesh Kumar via llvm-commits
- [llvm] [CostModel][X86] Improve vXi1 mask broadcast cost handling (PR #222467)
Dmitry Sidorov via llvm-commits
- [llvm] [NewPM] Increase inline size of IsResultInvalidated map (NFC) (PR #229367)
Florian Hahn via llvm-commits
- [llvm] [Instrumentor] Add atomicrmw instruction opportunity (PR #229369)
Ethan Luis McDonough via llvm-commits
- [clang] [llvm] [ARM] Fix musttail calls (PR #109943)
via llvm-commits
- [llvm] [AMDGPU][CFI] Replace CFA expression escapes with typed directives (PR #221011)
Alexander Hück via llvm-commits
- [llvm] [SPIRV] Splat scalar condition for vector OpSelect prior to SPIR-V 1.4 (PR #229267)
Arseniy Obolenskiy via llvm-commits
- [llvm] [GlobalISel][ARM] Enable extended LLT types (PR #229183)
Tim Gymnich via llvm-commits
- [llvm] [AllocToken] Add TypeFuncHash and TypeFuncHashPointerSplit modes (PR #227686)
Ai Nozaki via llvm-commits
- [flang] [llvm] [Flang] Introduce common::numeric_limits (PR #228108)
Eugene Epshteyn via llvm-commits
- [llvm] [IntrinsicLowering] Allow implicit truncation of ctpop masks (PR #229373)
Aochang Liu via llvm-commits
- [llvm] [X86] Prefer 128-bit vectors for Bulldozer CPUs (PR #225838)
Simon Pilgrim via llvm-commits
- [clang] [llvm] [mlir] [LLVMABI][AARCH64] Added aggregate return support (PR #229202)
Kunal Dubey via llvm-commits
- [llvm] [SLP] Avoid seeding related affine loop address computations (PR #226220)
Tim Besard via llvm-commits
- [llvm] [CGData] Fix section name lookup for COFF object files (PR #229375)
Ai Nozaki via llvm-commits
- [clang] [llvm] [mlir] [LLVMABI][AArch64] Support aggregate arguments (PR #229202)
Kunal Dubey via llvm-commits
- [llvm] 08483c5 - RISCV: Stop setting kill flags on virtual registers before FinalizeISel (#229013)
via llvm-commits
- [llvm] 0c53d0f - [X86] Require contract on FMUL while folding FMADDSUB to VFMULC (#229309)
via llvm-commits
- [llvm] [X86] Remove mul-constant-optimization flag (PR #229376)
Simon Pilgrim via llvm-commits
- [clang] [llvm] [TargetParser][clang] Parse and store Intel IGCA virtual architecture as an OffloadArch (PR #227468)
Nikita Kornev via llvm-commits
- [lld] [ELF] --gc-sections: keep a plain .gcc_except_table with its function (PR #229377)
via llvm-commits
- [llvm] [LV] Respect access type and alignment when speculating loads (PR #229379)
Dávid Bolvanský via llvm-commits
- [clang] [llvm] [AArch64] Implement the atomic fetch with hint intrinsic (PR #227711)
Kerry McLaughlin via llvm-commits
- [llvm] [Mips] Print ABI register names in MIR (PR #228017)
Alexander Richardson via llvm-commits
- [llvm] [SLP][NFC]Add an extra test with the reodering of the aggregate bitpacks, NFC (PR #229380)
Alexey Bataev via llvm-commits
- [llvm] [DebugInfo] Accept DW_OP_xderef_size and DW_OP_lit1..DW_OP_lit31. (PR #227839)
Matt Davis via llvm-commits
- [llvm] [AArch64][PAC] Reset `killed` operand flags in outlined functions (PR #221041)
Anatoly Trosinenko via llvm-commits
- [llvm] [VPlan] Mark vector-to-scalar VPInstructions as GeneratesSingleScalar (NFC). (PR #229381)
Florian Hahn via llvm-commits
- [llvm] [X86] Implement CRC32 const folding (PR #219452)
Simon Pilgrim via llvm-commits
- [llvm] [AMDGPU] Tolerate AMDGPUTargetELFStreamer without initializeTargetID (PR #226495)
Tim Renouf via llvm-commits
- [llvm] [X86] Fix constant vector i64 signed div/rem on 32-bit targets (PR #229244)
Rito Takeuchi via llvm-commits
- [llvm] [LoopVectorize] Loop-invariant backedge values prevent recognition of fixed-order recurrence (PR #228021)
via llvm-commits
- [llvm] [SystemZ][z/OS] Restore the frame header on stack restore (PR #228376)
MMS IT GmbH via llvm-commits
- [llvm] [SelectionDAG] Use correct result type in visitEXTRACT_VECTOR_ELT (PR #148707)
Simon Pilgrim via llvm-commits
- [llvm] [BOLT] Retain certain local symbols (PR #184074)
YongKang Zhu via llvm-commits
- [llvm] 1c694b9 - [AMDGPU] Measure MFMA hazard windows at their own producers (#225989)
via llvm-commits
- [llvm] [AMDGPU] Measure the remaining hazard windows at their own producers (PR #229167)
Janek van Oirschot via llvm-commits
- [llvm] [DAGCombiner] Forward a narrow load from an overlapping wider load (PR #212667)
Simon Pilgrim via llvm-commits
- [llvm] 7f0cd81 - [LV] Allow out-of-loop backedge users for min/max value of argmin/argmax. (#220351)
via llvm-commits
- [llvm] [SimplifyCFG] Do not convert branch to select for loop carried phi nodes (PR #227242)
Shreeyash Pandey via llvm-commits
- [llvm] [SystemZ] Custom lowering for vector min/max reductions (PR #229382)
via llvm-commits
- [llvm] [Offload][L0] Include context allocations in kernel indirect access flags (PR #229384)
Jan Trusiłło via llvm-commits
- [llvm] 27862ae - [GVN] Use willNotFreeBetween in loop-load PRE (#228033)
via llvm-commits
- [llvm] [X86] Keep scalar bf16/f16 selects in vector registers (PR #224218)
Simon Pilgrim via llvm-commits
- [llvm] c68c176 - [AArch64][PAC] Reset `killed` operand flags in outlined functions (#221041)
via llvm-commits
- [llvm] [ValueTracking] Compute known bits of and/or recurrences from start and step (PR #226164)
Ömer Sinan Ağacan via llvm-commits
- [llvm] 37ecc8c - [ValueTracking] Compute known bits of and/or recurrences from start and step (#226164)
via llvm-commits
- [llvm] [AMDGPU] Restrict DPP combine from performing bad transformations when handling certain REV insts that use Src1 as the DPP operand (PR #216835)
via llvm-commits
- [libc] [llvm] [libc] Fix math smoke powf_test's tolerance settings (PR #229390)
Victor Campos via llvm-commits
- [llvm] [BitcodeReader] Avoid integer overflow in strtab bounds check (PR #228786)
Prasad Deshpande via llvm-commits
- [llvm] [Mips] Add named COP0 registers (PR #229391)
Jiaxun Yang via llvm-commits
- [llvm] [X86] Skip debug instructions when adjusting the stack with pops (PR #229392)
Yongqiang Tian via llvm-commits
- [llvm] [SPIR-V] Add PreLegalizer pattern matching for ldexp (PR #227126)
Arseniy Obolenskiy via llvm-commits
- [llvm] [SelectionDAG] Avoid redundant comparisons in bit-test switch lowering (PR #222626)
Shreeyash Pandey via llvm-commits
- [llvm] [CGData] Fix reading unaligned stable function maps (PR #229394)
Ai Nozaki via llvm-commits
- [compiler-rt] [TySan] Make TySan compatible with sanitizer common interceptors (par… (PR #197688)
Matthew Nagy via llvm-commits
- [clang] [compiler-rt] [llvm] [TySan] Let TypeSanitizer know about conservative path TBAA data (PR #227782)
Matthew Nagy via llvm-commits
- [llvm] [SCEV] Strip stub known-false WrapPred check (PR #225086)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SCEV] Introduce exact flag for UDiv (PR #225065)
Ramkumar Ramachandra via llvm-commits
- [llvm] [Remarks][Support] Make YAML remark arguments round-trip (PR #227654)
Henrich Lauko via llvm-commits
- [llvm] [SystemZ][z/OS] Keep read-only data in the code section only if local and without relocations (PR #226683)
MMS IT GmbH via llvm-commits
- [llvm] [MachineCSE] Clear kill flags once per register in copy propagation (PR #226781)
MMS IT GmbH via llvm-commits
- [llvm] [AMDGPU] Explicit instruction prefetch for large entry functions (PR #227995)
via llvm-commits
- [llvm] [SLP]Support copyable GEP/zext main ops and relax gather pointer compatibility (PR #221599)
Hans Wennborg via llvm-commits
- [llvm] Revert "[SLP]Support copyable GEP/zext main ops and relax gather pointer compatibility" (PR #229397)
Hans Wennborg via llvm-commits
- [clang] [llvm] [IPO] Add InternalizeLinkOnceODR pass to enable IPO on linkonce_odr functions (PR #229402)
Kuba Mracek via llvm-commits
- [llvm] [BOLT] Update retained internal symbol addresses (PR #229407)
Adam Bzowski via llvm-commits
- [llvm] [DWARF] Don't index variables whose address is only their value (PR #228574)
LLVM Continuous Integration via llvm-commits
- [llvm] [LoopUnroll] Relax fixed non-latch exit probability threshold for runtime multi-exit loop unrolling (PR #224325)
Marek Sedláček via llvm-commits
- [llvm] [NVPTX] Infer address spaces for mbarrier intrinsics (PR #229272)
Durgadoss R via llvm-commits
- [llvm] [SelectionDAG] Promote approximate f16 vector FMA to f32 (PR #228900)
KIM MYEONG SEOK via llvm-commits
- [llvm] [CostModel] Generic cost model partial reduce (PR #229426)
Graham Hunter via llvm-commits
- [llvm] [AMDGPU] Extend new hazard CFG walk for wmma instruction support (PR #229428)
Janek van Oirschot via llvm-commits
- [llvm] [AMDGPU] Fix long-branch jump clobbers callee-saved registers (PR #229427)
Jan Patrick Lehr via llvm-commits
- [clang] [llvm] [AMDGPU[ Expand `av` built-ins and intrinsics to b16/b32/b64 (PR #229429)
Pierre van Houtryve via llvm-commits
- [llvm] 409b1cd - [GlobalISel] Add GIHasOneUse builtin for combiner match patterns (#221804)
via llvm-commits
- [llvm] [Github] Bump runner version to 2.338.0 (PR #229439)
Aiden Grossman via llvm-commits
- [clang] [llvm] [mlir] [polly] [IRBuilder] Accept Module instead of LLVMContext in ctor (PR #229430)
Nikita Popov via llvm-commits
- [llvm] [ScheduleDAG](NFC) Allocate worklist once in `ScheduleDAGTopologicalSort` (PR #179256)
Florian Hahn via llvm-commits
- [llvm] [CodeGen] Ignore non-integer unsafe-stack-size annotation (PR #229440)
Nikita Popov via llvm-commits
- [llvm] [X86] Don't skip frame setup/destroy on SEH pseudos and regenerate tests (PR #225445)
LLVM Continuous Integration via llvm-commits
- [llvm] [KnownFPClass] Fix KnownFPClass::frem subnormal handling and improve deductions. (PR #219303)
Eric Ross via llvm-commits
- [llvm] [AMDGPU][NewInsertWaitcnt] Add Event Tracker (PR #226970)
Pierre van Houtryve via llvm-commits
- [llvm] [VPlan] Run CSE in VPlanTransforms::optimize (PR #203370)
Ramkumar Ramachandra via llvm-commits
- [llvm] [NVPTX] Use register-or-immediate operands for fns (PR #229224)
Alex MacLean via llvm-commits
- [llvm] [AMDGPU] Don't apply the memory cluster budget to hard clause formation (PR #229445)
Barbara Mitic via llvm-commits
- [llvm] AMDGPU: Use HasUsableDSOffset for LDS read2/write2 patterns (PR #229446)
Matt Arsenault via llvm-commits
- [llvm] [SimplifyCFG] Skip phi scans when folded block shares no predecessors (PR #225647)
via llvm-commits
- [llvm] [SPIRV] Don't cache SyncScope IDs across LLVMContexts in getMemScope (PR #229444)
Tim Besard via llvm-commits
- [llvm] Revert VectorCombine (PR #229449)
Vitaly Buka via llvm-commits
- [llvm] [InstCombine] Fix profiles in EvaluateInDifferentTypeImpl (PR #229447)
Mircea Trofin via llvm-commits
- [llvm] [GISel][AArch64][Mips][RISCV] Allow G_BRINDIRECT to be selected via tablegen. (PR #229450)
Craig Topper via llvm-commits
- [llvm] 681e59f - [LiveDebugVariables] Repair stale SlotIndexes (#224467)
via llvm-commits
- [llvm] [X86] Reuse SUB results for conditional subtract (PR #229452)
Hadong Lee via llvm-commits
- [llvm] [NFC][ADT] Replace memcpy with range-based copies (PR #229451)
Younes Yazami via llvm-commits
- [compiler-rt] a63780f - [NFC][TSan] Use ScopedReport::AddThread(Tid) in ReportRace (#228813)
via llvm-commits
- [llvm] [JumpThreading] Fix assert during compilation, bug #229448 (PR #229465)
via llvm-commits
- [compiler-rt] [NFC][TSan] Use ScopedReport::AddThread(Tid) in ReportRace (PR #228813)
Vitaly Buka via llvm-commits
- [llvm] [llubi] Support the oracle form of llvm.speculative.load. (PR #227095)
Florian Hahn via llvm-commits
- [llvm] 52281ef - [VPlan] Run CSE in VPlanTransforms::optimize (#203370)
via llvm-commits
- [llvm] c9e13e0 - [SystemZ][z/OS] Keep read-only data in the code section only if local and without relocations (#226683)
via llvm-commits
- [llvm] dc17218 - AMDGPU: Use HasUsableDSOffset for LDS read2/write2 patterns (#229446)
via llvm-commits
- [llvm] [InstCombine] Propagate demanded elements through trivially-vectorizable intrinsics (PR #221495)
Hugo Pompougnac via llvm-commits
- [llvm] [IndirectBrExpand] Preserve profile weights (PR #227784)
Mircea Trofin via llvm-commits
- [compiler-rt] [NFC][TSan] Lock ThreadRegistryLock inside ScopedReport::AddThread and AddLocation (PR #228814)
Vitaly Buka via llvm-commits
- [llvm] [X86][CostModel] Price predicate mask-expansion fanout for wide predicated ops (PR #201666)
via llvm-commits
- [llvm] [LiveDebugVariables] Add tests for stale SlotIndexes (PR #225375)
Jeremy Morse via llvm-commits
- [compiler-rt] [Hexagon] Add v79 QFloat HVX runtime conversion funcs (PR #229481)
Fateme Hosseini via llvm-commits
- [clang] [llvm] [Clang][AArch64] Prevent hotpatching from intercepting loop backedges (PR #229477)
Eli Friedman via llvm-commits
- [llvm] [X86][TTI]Model the sqrt folded into the reciprocal estimate as free (PR #229482)
Alexey Bataev via llvm-commits
- [llvm] 4e3d729 - Revert [VectorCombine] Combine adjacent loads feeding as shuffle operands #213007 (#229449)
via llvm-commits
- [compiler-rt] [NFC][TSan] Remove redundant ThreadRegistryLock debug check in ShouldReport (PR #228815)
Vitaly Buka via llvm-commits
- [llvm] [Hexagon][NFC] Fix duplicate A2_tfr check in HexagonGlobalScheduler (PR #227762)
Fateme Hosseini via llvm-commits
- [llvm] 0f967eb - [FastISel] Ignore debug uses when folding a load (#228719)
via llvm-commits
- [llvm] AMDGPU: Select region address space loads and stores (PR #229500)
Matt Arsenault via llvm-commits
- [llvm] b30c031 - [SeparateConstOffsetFromGEP] Fix sext/add overflow checks (#228562)
via llvm-commits
- [llvm] Fix SystemZ TableGen crashing when GlobalISel is enabled (PR #228227)
Ulrich Weigand via llvm-commits
- [llvm] RuntimeLibcallsEmitter: Handle calling conv in per-library functions (PR #223760)
Paul Kirth via llvm-commits
- [llvm] [AMDGPU] Use vector memory types for 64/128-bit cooperative atomics (PR #229501)
Gheorghe-Teodor Bercea via llvm-commits
- [llvm] [InstCombine] Simplify demanded elements of vector intrinsics (PR #229498)
Rudrajeet Pal via llvm-commits
- [compiler-rt] [RISCV][compiler-rt] Asm funcs should be hidden (PR #226633)
Sam Elliott via llvm-commits
- [llvm] [AMDGPU] Omit hardwired-on SRAMECC ELF mode (PR #228268)
Krzysztof Drewniak via llvm-commits
- [llvm] f28445b - [SDPatternMatch] Add m_Node and m_SpecificOpc that take opcode from template argument (#228626)
via llvm-commits
- [llvm] [NVPTX] Add L1 32B prefetch intrinsics (PR #227809)
Alex MacLean via llvm-commits
- [llvm] [GlobalISel] Ignore debug instructions in trunc-store search budget (#224660) (PR #229492)
Aditya Roy Chowdhury via llvm-commits
- [compiler-rt] ccb7fb8 - [RISCV][compiler-rt] Asm funcs should be hidden (#226633)
via llvm-commits
- [llvm] [NFC][ARM] Remove unneeded cast (PR #229504)
Ben Wu via llvm-commits
- [llvm] 352a92f - Fix SystemZ TableGen crashing when GlobalISel is enabled (#228227)
via llvm-commits
- [llvm] [CI] Enable CIR testing for all MLIR and Clang changes (PR #227834)
Aaron Ballman via llvm-commits
- [llvm] [AVR] Fix a crash when generating debug information for frame pointer (PR #227285)
Patryk Wychowaniec via llvm-commits
- [llvm] [GlobalISel] Ignore debug instructions in trunc-store search budget (PR #229492)
Aditya Roy Chowdhury via llvm-commits
- [llvm] Revert "[VPlan] Run CSE in VPlanTransforms::optimize" (PR #229507)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SPIR-V] Add OpUntypedVariableLengthArrayINTEL support for SPV_INTEL_variable_length_array (PR #224756)
Vyacheslav Levytskyy via llvm-commits
- [lld] [llvm] [dyndbg][LLD][ELF] Add dynamic debugging support for `--wrap=<symbol>` (PR #222990)
Andrew Ng via llvm-commits
- [llvm] [AMDGPU] Account for gfx950 permlane hazards during scheduling (PR #229517)
Alan Li via llvm-commits
- [llvm] [InferAddressSpaces] Do not rewrite index operands of a constant GEP (PR #229491)
Malavika Samak via llvm-commits
- [llvm] [InstCombine] Fold Carry-Save Adder identities: `(A ^ B) + ((A & B) << 1)` to `A + B` and `(A + B) - ((A & B) << 1)` to `A ^ B` (PR #229511)
Nikita Popov via llvm-commits
- [llvm] [SDag] Notify listeners when deleting a node (PR #66991)
Sergei Barannikov via llvm-commits
- [llvm] f9cdffa - AMDGPU: Select region address space loads and stores (#229500)
via llvm-commits
- [llvm] [GlobalISel] Preserve cache hint metadata through IRTranslator (PR #224522)
Matt Arsenault via llvm-commits
- [llvm] 2e091b2 - [NewPM] Properly initialize NewPM reg-bank-select (#229534)
via llvm-commits
- [llvm] [NewPM] Properly initialize NewPM reg-bank-select (PR #229534)
Mikołaj Piróg via llvm-commits
- [flang] [llvm] [mlir] Revert "[flang][mlir][OpenMP] Report Fortran names for privatized target maps" (PR #229549)
Jason Van Beusekom via llvm-commits
- [llvm] AMDGPU/GlobalISel: Split 4-byte aligned 64-bit LDS accesses on SI (PR #229531)
Dmitry Sidorov via llvm-commits
- [llvm] b488848 - [SPIRV] Splat scalar condition for vector OpSelect prior to SPIR-V 1.4 (#229267)
via llvm-commits
- [llvm] [DirectX] Cast the pointer for atomics whose value type differs (PR #226563)
Joshua Batista via llvm-commits
- [clang] [llvm] [clang] Deduplicate builtin diagnostic descriptions (PR #202624)
Kelvin Li via llvm-commits
- [llvm] RegisterCoalescer: Restore subrange PHI inputs of pruned undef values (PR #229551)
Matt Arsenault via llvm-commits
- [clang] [llvm] [Inliner] Use a disjoint scope domain for noalias arguments (PR #218769)
Krzysztof Drewniak via llvm-commits
- [lld] [lld][WebAssembly] Error on TLS relocations against shared library symbols (PR #229464)
Sam Clegg via llvm-commits
- [llvm] 53a26d0 - AMDGPU/GlobalISel: Split 4-byte aligned 64-bit LDS accesses on SI (#229531)
via llvm-commits
- [clang] [llvm] [clang][Driver] Support Trailing Comments in Config Files (PR #213765)
Sam Elliott via llvm-commits
- [llvm] [ConstraintElimination] Support widened (zext/sext) induction variables (PR #223214)
Florian Hahn via llvm-commits
- [llvm] [VectorCombine] Add new instructions generated by deinterleave2 folding into worklist (PR #229558)
Sam Elliott via llvm-commits
- [compiler-rt] [sanitizer_common] Log a backtrace for unexpected format printf specifiers (PR #222658)
Dan Blackwell via llvm-commits
- [clang] [llvm] [AArch64] Prevent hotpatching from intercepting loop backedges (PR #229477)
Alexandre Ganea via llvm-commits
- [llvm] [AMDGPU] Fix DS result latency for inline asm (PR #229538)
Alan Li via llvm-commits
- [llvm] [IR] Add an "atomicity" operand bundle (PR #223298)
Krzysztof Drewniak via llvm-commits
- [llvm] [X86] Mark .lrodata for non-mergeable constants with SHF_X86_64_LARGE (PR #229279)
Grigory Pastukhov via llvm-commits
- [clang] [llvm] Implement WaveGetLaneCount SPIRV lowering (PR #227160)
Joshua Batista via llvm-commits
- [llvm] 069eb04 - [AMDGPU] Account for gfx950 permlane hazards during scheduling (#229517)
via llvm-commits
- [llvm] [LV] Thread an explicit CM param through VPlan-building/costing helpers (PR #224836)
Florian Hahn via llvm-commits
- [llvm] [AMDGPU] Model gfx1250 f8f6f4 WMMA latency in llvm-mca (PR #227150)
via llvm-commits
- [flang] [llvm] [flang-rt][cuda] Add CUFRegisterHostMemoryRange entry point (PR #229572)
Zhen Wang via llvm-commits
- [llvm] [LV] Use frozen start values from the main plan directly (NFC). (PR #229571)
Florian Hahn via llvm-commits
- [llvm] [RISCV] Fold (sub 0, (srl (and X, (1 << ShAmt)), ShAmt)) -> (sra (shl X, ShAmt2), bits-1) (PR #229509)
Craig Topper via llvm-commits
- [clang] [llvm] [SPIR-V] Legalize byte-address buffer atomic operations (PR #223874)
Joshua Batista via llvm-commits
- [llvm] [Bazel] Add FlowSensitive/Models headers to clang:analysis (PR #229578)
Anshul Nigham via llvm-commits
- [clang] [llvm] [HLSL] Add Barrier builtin (PR #228277)
Deric C. via llvm-commits
- [llvm] [RISCV] Fix violations of SDTCisSameAs when creating vector ISD nodes. (PR #228215)
Craig Topper via llvm-commits
- [clang] [llvm] [InferAttrs] Avoid conflicting argument memory attributes (PR #226207)
via llvm-commits
- [llvm] [GVN] Allow loop-load PRE when the blocker is in an inner loop (PR #229580)
Anup Kumar via llvm-commits
- [llvm] [Option] Allow subcommand names as positional arguments (PR #227077)
Prabhu Rajasekaran via llvm-commits
- [llvm] [X86] Fix LDTILECFG insertion at block starts (PR #229239)
Yongqiang Tian via llvm-commits
- [llvm] [LICM] Recompute insertion point for promotion when necessary (PR #229581)
Eli Friedman via llvm-commits
- [llvm] [RISCV] Use ComplexPattern for matching addi+brind. (PR #229594)
Matt Arsenault via llvm-commits
- [llvm] [llvm-c][ocaml] Move DataLayout bindings from Target to Core/IR (PR #227714)
Sergei Barannikov via llvm-commits
- [llvm] [Bazel][libc] Update platform_file to depend on dup3 (PR #229595)
Anshul Nigham via llvm-commits
- [clang] [llvm] [msan] Don't instrument llvm::fake_use (PR #229591)
Thurston Dang via llvm-commits
- [llvm] [docs][llvm-readobj] Document the --call-graph-info output format (PR #227386)
Prabhu Rajasekaran via llvm-commits
- [llvm] [LAA] Allow runtime checks between comparable address spaces. (PR #229588)
Matt Arsenault via llvm-commits
- [llvm] [BOLT] Match mmap events by build id (PR #229596)
Fabian Parzefall via llvm-commits
- [clang] [llvm] [msan] Silently ignore llvm::fake_use (PR #229591)
Thurston Dang via llvm-commits
- [llvm] [SelectionDAG] Add ATOMIC_LOAD_FMAXIMUMNUM/FMINIMUMNUM to SDNode::getOperationName. (PR #229566)
Craig Topper via llvm-commits
- [llvm] [SelectionDAG] Add static_assert for ISD::BUILTIN_OP_END to SDNode::getOperationName to encourage updating when opcodes are added. (PR #229573)
Craig Topper via llvm-commits
- [llvm] [Bazel] Add Support dependency to amdgpu_tests (PR #229597)
Anshul Nigham via llvm-commits
- [llvm] [CAS][unittests] Relax time out for ProgramTest (PR #229598)
Steven Wu via llvm-commits
- [llvm] [AMDGPU] Fix producer latency for inline asm (PR #229538)
Stanislav Mekhanoshin via llvm-commits
- [llvm] [InstCombine] Treat `asin`, `asinh`, `atan` and `cbrt` as odd-functions (PR #227336)
via llvm-commits
- [clang] [llvm] [docs] Drop stale "experimental" note for preserve_most (PR #214205)
John McCall via llvm-commits
- [clang] [llvm] [msan] Correctly handle llvm::fake_use (as a no-op) (PR #229591)
Thurston Dang via llvm-commits
- [libc] [llvm] [libc][math] Improve accuracy for exp*f float-only implementations. (PR #227881)
via llvm-commits
- [llvm] [SelectionDAG] Preserve cache hint metadata through DAGCombiner (PR #225954)
Yonah Goldberg via llvm-commits
- [llvm] 92dfb7a - [msan] Correctly handle llvm::fake_use (as a no-op) (#229591)
via llvm-commits
- [clang] [flang] [libc] [libcxx] [lldb] [llvm] [mlir] [GlobalISel] Preserve cache hint metadata through IRTranslator (PR #224522)
Yonah Goldberg via llvm-commits
- [llvm] [CycleAnalysis] Add missing export to symbol used in test (PR #228316)
Tom Stellard via llvm-commits
- [llvm] [PredicateInfo] Support select instructions (PR #229613)
Jon Roelofs via llvm-commits
- [llvm] workflows/precommit: Use single shared job to compute projects (PR #224941)
Tom Stellard via llvm-commits
- [clang] [flang] [llvm] [CodeGen] Add a FunctionSplittingMode target option (PR #226575)
Sriraman Tallam via llvm-commits
- [llvm] [AsmPrinter] Fix direct callee operand lookup in CallGraphSection (PR #218534)
Prabhu Rajasekaran via llvm-commits
- [llvm] [RISCV][NFC] Simplify isSupportedStackID (PR #229117)
Sudharsan Veeravalli via llvm-commits
- [llvm] 4a5b1cd - [RISCV][NFC] Simplify isSupportedStackID (#229117)
via llvm-commits
- [llvm] Improve DWARFDebugLine::LineTable::lookupAddressRangeImpl(...) to act as exptected. (PR #226625)
Greg Clayton via llvm-commits
- [llvm] 03302cd - [RISCV] Add Xqcisls loads to getJumpTableIndex (#229006)
via llvm-commits
- [lld] [lld][macho] Fix STABS for branch extension thunks with ICF (PR #229604)
Ellis Hoag via llvm-commits
- [llvm] [DebugInfo] Preserve irlayers across inlining, remapping and hoisting (PR #229615)
Alexander Yermolovich via llvm-commits
- [llvm] 13d5a13 - [AsmPrinter] Fix direct callee operand lookup in CallGraphSection (#218534)
via llvm-commits
- [compiler-rt] [llvm] [sanitizer_common] Intercept fortified __mem*_chk functions (PR #206702)
Vitaly Buka via llvm-commits
- [llvm] [AMDGPU] Fix missed WMMA C-operand co-exec hazard (PR #226503)
Petar Jovanovic via llvm-commits
- [llvm] [MIR] Serialize the "foldable" inline asm register operand flag (PR #229626)
Bill Wendling via llvm-commits
- [llvm] AMDGPU: VALU data fast-forwarding needs no s_delay (PR #205481)
Reem Elkhouly via llvm-commits
- [llvm] [ThinLTO] Only replace uses in ExportM when promoting CFI functions (PR #229634)
Mircea Trofin via llvm-commits
- [llvm] [InstCombine] Fix profile propagation in foldNestedSelects (PR #229635)
Aiden Grossman via llvm-commits
- [llvm] [DirectX] Generalize DXIL integer legalization (PR #225233)
Farzon Lotfi via llvm-commits
- [llvm] [MLGO][Docs] Document MLIR based model embedding (PR #227536)
Mircea Trofin via llvm-commits
- [llvm] [AMDGPU][GIsel] Properly handle unsupported intrinsic translation (PR #228040)
Vikram Hegde via llvm-commits
- [flang] [llvm] [flang-rt][NFC] Fix too short first lines (PR #229642)
Valentin Clement バレンタイン クレメン via llvm-commits
- [llvm] [KnownFPClass] Apply denormal modes at the fmul/fdiv boundaries (PR #229638)
Eric Ross via llvm-commits
- [llvm] [RISCV] Add a RISCVISD::TUPLE_CAST to represent casting one tuple to another. (PR #228249)
Brandon Wu via llvm-commits
- [llvm] [IR] Name unnamed values after each pass (PR #221519)
Arthur Eubanks via llvm-commits
- [llvm] [msan] Handle vector.partial.reduce.{add,fadd} (PR #229649)
Thurston Dang via llvm-commits
- [llvm] [DataLayout] Add byte specification (PR #106536)
Sergei Barannikov via llvm-commits
- [llvm] 8cee9e8 - [InstCombine] Preserve profile metadata when folding fneg selects (#228754)
via llvm-commits
- [llvm] [InstCombine] Preserve profile metadata when folding fneg selects (PR #228754)
Aiden Grossman via llvm-commits
- [llvm] [AArch64] Add separate Firestorm scheduling model (PR #229641)
Bardi Harborow via llvm-commits
- [llvm] [InstCombine] Fix profiles in sinkNotIntoOtherHandOfLogicalOp (PR #229555)
Aiden Grossman via llvm-commits
- [llvm] [MLGO] Switch regalloc log context once per function (PR #229453)
Arseniy Obolenskiy via llvm-commits
- [llvm] [KnownFPClass] Correct subnormal handling in ldexp (PR #225904)
Eric Ross via llvm-commits
- [llvm] a129b3c - [KnownFPClass] Correct subnormal handling in ldexp (#225904)
via llvm-commits
- [llvm] [NFC][X86] Simplify target assert (PR #229643)
Ben Wu via llvm-commits
- [llvm] [SLP] Store reduction context hint as part of VectorValuesAndScales (PR #229653)
Ryan Buchner via llvm-commits
- [llvm] [SLP] Pass along stride information when costing strided loads and stores (PR #227823)
Ryan Buchner via llvm-commits
- [llvm] [SelectionDAG] Reduce RRList scheduling overhead (PR #223241)
via llvm-commits
- [llvm] [IndVarSimplify] Skip exit predication past divergent effects (#183906) (PR #194202)
Jake Xue via llvm-commits
- [llvm] 7eea74b - [RISCV] Add a RISCVISD::TUPLE_CAST to represent casting one tuple to another. (#228249)
via llvm-commits
- [llvm] c01b660 - [MachineScheduler] Add support for scheduling while in SSA (#161054)
via llvm-commits
- [llvm] [MachineScheduler] Add support for scheduling while in SSA (PR #161054)
Carl Ritson via llvm-commits
- [llvm] 161d9dc - [SimplifyLibCalls] Fix profile propagation in replacePowWithSqrt (#229645)
via llvm-commits
- [llvm] [SimplifyLibCalls] Fix profile propagation in replacePowWithSqrt (PR #229645)
Aiden Grossman via llvm-commits
- [llvm] [SandboxVec][LoadStoreVec] Support constant vectors of mixed types (PR #224890)
Anshil Gandhi via llvm-commits
- [llvm] 0fe07d8 - [InstCombine] Fix profiles in sinkNotIntoOtherHandOfLogicalOp (#229555)
via llvm-commits
- [llvm] [AArch64] Implement CRC32 const folding (PR #228985)
Shreesh Adiga via llvm-commits
- [llvm] [RegAllocFast] Do not free the register of an error-assigned def (PR #229657)
Aochang Liu via llvm-commits
- [llvm] [InstCombine] Fix InstCombine pass to sink llvm.assume calls (PR #229658)
via llvm-commits
- [libc] [llvm] [libc][math] Optimize integer-only expf and add integer-only exp2f and exp10f (PR #228898)
via llvm-commits
- [llvm] c64ebf9 - [SLP] Pass along stride information when costing strided loads and stores (#227823)
via llvm-commits
- [clang] [compiler-rt] [ubsan] Improve scoped enum diagnostic printer (PR #216460)
via llvm-commits
- [llvm] [Hexagon] Do not promote unaligned loop access to memmove. (PR #223936)
via llvm-commits
- [llvm] Reland "[VectorCombine] Combine adjacent loads feeding as shuffle operands" (PR #229662)
Sushant Gokhale via llvm-commits
- [llvm] [orc-rt] Mark ControllerAccess constructor as noexcept. (PR #229663)
Lang Hames via llvm-commits
- [compiler-rt] [tsan] Handle MAP_FIXED_NOREPLACE as an exact mmap request (PR #229603)
Vitaly Buka via llvm-commits
- [llvm] [TwoAddressInstruction] Don't mark a use undef if it reads a defined lane (PR #229554)
Vitaly Buka via llvm-commits
- [compiler-rt] [scudo] Validate list endpoints and links before removing a node (PR #229245)
Vitaly Buka via llvm-commits
- [llvm] [LLVM][NVPTX] Add fabric.try_put intrinsic support (PR #229665)
Rajat Bajpai via llvm-commits
- [llvm] [examples] Set target triple on modules in Kaleidoscope. (PR #229666)
Lang Hames via llvm-commits
- [llvm] [AMDGPU][MLGO] Add regalloc scoring pass to the optimized pipeline (PR #229668)
Arseniy Obolenskiy via llvm-commits
- [llvm] [ARM] Add vtrn/vzip/vuzp shuffle costs. (PR #223611)
David Green via llvm-commits
- [llvm] [SPARC] Implement TargetInstrInfo::insertIndirectBranch() (PR #226678)
Daniel Cederman via llvm-commits
- [llvm] [X86] Relax SDTypeProfile for X86ISD::CALL to allow i64 pointers for X32 (PR #229670)
Craig Topper via llvm-commits
- [llvm] [LoopVectorize] Don't speculate an early-exit trip count that may cause UB (PR #219893)
Madhur Amilkanthwar via llvm-commits
- [llvm] [AMDGPU] Ignore debug instructions in GPR-index scan distance (PR #224279)
Aochang Liu via llvm-commits
- [llvm] [AMDGPU][Codegen] Legalize 16bit register class in sdag (PR #220275)
Guo Chen via llvm-commits
- [llvm] [X86] Fix equality operator ambiguity (PR #152739)
Gregor Jasny via llvm-commits
- [llvm] [NVVM][NVPTX] Support spcompress and spdecompress intrinsics (PR #221751)
Kirill Vedernikov via llvm-commits
- [llvm] AMDGPU/GlobalISel: Use integer types when narrowing loads and stores (PR #229547)
Matt Arsenault via llvm-commits
- [llvm] [RegAllocFast] Allocate tied uses first and enable for PowerPC (PR #229672)
Fangrui Song via llvm-commits
- [llvm] AMDGPU/GlobalISel: Bitcasting G_TRUNC combine (PR #217634)
Matt Arsenault via llvm-commits
- [llvm] 770a601 - [ARM] Add vtrn/vzip/vuzp shuffle costs. (#223611)
via llvm-commits
- [llvm] [SPARC] Lower dynamic v2i32 vector extracts on 32-bit SPARC (PR #222348)
Daniel Cederman via llvm-commits
- [llvm] [Loads] Add MUL/SHL handling in decomposeLinearExpression (PR #222574)
Arne Stenkrona via llvm-commits
- [llvm] [AMDGPU] Add module flag for LDS size limit (PR #229557)
Mariusz Sikora via llvm-commits
- [llvm] [offload][omp] Route data memory operations through liboffload (PR #226443)
Alex Duran via llvm-commits
- [llvm] [SPIRV] Re-sort basic blocks after LoopSimplify for non-shader targets (PR #229678)
Alexander Shaposhnikov via llvm-commits
- [llvm] [OCaml] Remove unused open Llvm_target in test (NFC) (PR #229680)
Nikita Popov via llvm-commits
- [llvm] AArch64: Test trap-after-noreturn against the exception model (PR #229681)
Matt Arsenault via llvm-commits
- [llvm] CodeGen: Remove TRI arguments from TargetInstrInfo hooks (PR #228164)
Matt Arsenault via llvm-commits
- [llvm] [RISCV] - Override generic DefaultUnrollRuntimeCount value for RISCV to 4. (PR #229684)
Pawan Nirpal via llvm-commits
- [llvm] AMDGPU/GlobalISel: Fix buffer stores of bf16 (PR #229673)
Anshil Gandhi via llvm-commits
- [llvm] [flang] Blank-fill internal list-directed record when no output is emitted (PR #213205)
via llvm-commits
- [llvm] b7fedff - [GlobalISel] Ensure integer trunc and anyext in call lowering. (#229346)
via llvm-commits
- [llvm] [flang-rt] Split large unformatted sequential records into subrecords (PR #213003)
via llvm-commits
- [llvm] ee22ca5 - AMDGPU/GlobalISel: Fix buffer stores of bf16 (#229673)
via llvm-commits
- [llvm] 2cc7a05 - [MachineCombiner] Don't count debug instructions in block size (#225765)
via llvm-commits
- [clang] [llvm] [AArch64] Codegen for AArch64 Return Address Authentication Hardening (PR #176187)
Anatoly Trosinenko via llvm-commits
- [llvm] [Analysis] Remove unused CallGraphSCC methods (NFC) (PR #229683)
Nikita Popov via llvm-commits
- [llvm] TableGen: Allow IsTruncStore predicates on atomic PatFrags (PR #229687)
Matt Arsenault via llvm-commits
- [llvm] AMDGPU: Fix selection of atomicrmw usub_sat for LDS and flat (PR #229471)
Tim Besard via llvm-commits
- [llvm] f4b6e8c - [SPIRV] Re-sort basic blocks after LoopSimplify for non-shader targets (#229678)
via llvm-commits
- [polly] ecc201f - [IRBuilder] Accept Module instead of LLVMContext in ctor (#229430)
via llvm-commits
- [llvm] Mips: Mark the assembler temporary def dead in inserted branches (PR #229690)
Matt Arsenault via llvm-commits
- [llvm] [SCEV] Do not consider base pointer defining scope for offset flags in `getGEPExpr` (PR #222713)
Antonio Frighetto via llvm-commits
- [llvm] 16792e0 - [OCaml] Remove unused open Llvm_target in test (NFC) (#229680)
via llvm-commits
- [llvm] Revert "[IndirectBrExpand] Preserve profile weights" (PR #229691)
Nikita Popov via llvm-commits
- [polly] [Polly] Fall back to the exact dependences if scheduling fails (PR #229692)
Timur Baidusenov via llvm-commits
- [llvm] 389a69b - [Analysis] Remove unused CallGraphSCC methods (NFC) (#229683)
via llvm-commits
- [llvm] [JumpThreading] Do not fold compare when incoming PHI value is defined in current block (#226049) (PR #226247)
Divyansh Yadav via llvm-commits
- [llvm] e128a79 - [AArch64] Add predicate-as-counter loop rewrite pass (#220960)
via llvm-commits
- [llvm] [AArch64] Add predicate-as-counter loop rewrite pass (PR #220960)
Benjamin Maxwell via llvm-commits
- [llvm] [AArch64] Avoid materializing full masks for extractelement users of predicate-as-counter (PR #220961)
Benjamin Maxwell via llvm-commits
- [llvm] [X86] Fix failure to lower i8/i16 bzhi pattern to bzhi (PR #226401)
Divyansh Yadav via llvm-commits
- [llvm] [Offload][omp] Use olGetDeviceInfo to get device info (PR #229693)
Alex Duran via llvm-commits
- [clang] [clang-tools-extra] [flang] [libcxx] [llvm] [docs] Update obsolete Phabricator references for GitHub workflows (PR #229542)
Nikolas Klauser via llvm-commits
- [llvm] [llubi] Fix floating-point denormal flushing behavior (PR #229698)
Zhige Chen via llvm-commits
- [llvm] Targets: Remove redundant TRI arguments from InstrInfo helpers (PR #228165)
Matt Arsenault via llvm-commits
- [clang] [llvm] [polly] [IRBuilder] Make IRBuilder inserter non-virtual (PR #229700)
Nikita Popov via llvm-commits
- [llvm] [llvm-readobj][ELF] Support SHT_REL/SHT_RELA relocations for --call-graph-info on ET_REL (PR #213156)
James Henderson via llvm-commits
- [llvm] 3378a49 - AMDGPU: Fix selection of atomicrmw usub_sat for LDS and flat (#229471)
via llvm-commits
- [llvm] AArch64/GlobalISel: Mark ptrauth pseudo clobbers dead (PR #229701)
Matt Arsenault via llvm-commits
- [llvm] fix: recalculate .sxdata post strip (PR #222929)
Emily Sequeira via llvm-commits
- [llvm] [llubi] Validate argument types in `printf` (PR #229702)
Zhige Chen via llvm-commits
- [llvm] [X86] Ignore debug instructions when choosing LVI graph leaders (PR #224496)
Orlando Cazalet-Hyams via llvm-commits
- [clang] [llvm] Disentangle DISubprogram ODR type method uniquing from metadata uniquing (PR #217042)
Orlando Cazalet-Hyams via llvm-commits
- [llvm] [Offload] Avoid global mutex destructors (PR #221471)
Vedran Miletić via llvm-commits
- [llvm] [WebAssembly] Fix chunk order when reconstructing illegal bitmasks (PR #226824)
via llvm-commits
- [llvm] e699478 - Revert "[IndirectBrExpand] Preserve profile weights" (#229691)
via llvm-commits
- [llvm] [ARM][Thumb1] Account for tail duplication in fn size estimation (PR #214228)
Simon Tatham via llvm-commits
- [llvm] f549451 - RuntimeLibcallsEmitter: Add isolated LibcallLibrary variant (#223764)
via llvm-commits
- [llvm] RuntimeLibcallsEmitter: Add isolated LibcallLibrary variant (PR #223764)
Matt Arsenault via llvm-commits
- [llvm] RuntimeLibcalls: Dedup dispatch calls to the same library function (PR #229561)
Matt Arsenault via llvm-commits
- [llvm] [X86] Prefer v8i64 MULHS expansion on non-DQI AVX512 targets (PR #229704)
Simon Pilgrim via llvm-commits
- [llvm] AMDGPU: Apply xnack/sramecc settings before constructing TargetLowering (PR #229705)
Matt Arsenault via llvm-commits
- [llvm] 708fee3 - [AArch64][SME] Allow more inlining when SME attributes are incompatible. (#223393)
via llvm-commits
- [llvm] [SLSR] Guard variable delta reuse with canReuseInstruction (PR #225377)
Adrian Kuegel via llvm-commits
- [compiler-rt] 8ea9a49 - [TySan] Make TySan compatible with sanitizer common interceptors (par… (#197688)
via llvm-commits
- [llvm] [Offload][omp] Add own Error class to libomptarget (PR #229711)
Alex Duran via llvm-commits
- [llvm] [X86] Make EXTRQI/INSERTQI SDTypeProfile accurate and enable verifyTargetNode (PR #229655)
Simon Pilgrim via llvm-commits
- [llvm] [BOLT][AArch64] Preserve tail-call annotation on local trampolines (PR #228105)
Alexandros Lamprineas via llvm-commits
- [llvm] [LLVM][CodeGen][SVE] Add lowering for bfloat strict-fp cast operators. (PR #223709)
Paul Walker via llvm-commits
- [llvm] [flang] Fix value separators for CHARACTER array output (PR #205303)
via llvm-commits
- [llvm] 445981d - RegisterCoalescer: Restore subrange PHI inputs of pruned undef values (#229551)
via llvm-commits
- [flang] [llvm] [flang][MIF] Adding lowering for EVENT POST/WAIT statements (PR #210283)
Jean-Didier PAILLEUX via llvm-commits
- [llvm] 7348e71 - [AMDGPU][GIsel] Properly handle unsupported intrinsic translation (#228040)
via llvm-commits
- [llvm] [X86] Rework isel for BRIND for X32 to fix verifyTargetNode error. NFCI (PR #229661)
Simon Pilgrim via llvm-commits
- [llvm] c198ce3 - Targets: Remove redundant TRI arguments from InstrInfo helpers (#228165)
via llvm-commits
- [llvm] [SLP] Add store-to-load forwarding conflict cost for widened store chains (PR #199606)
Milin Bhade via llvm-commits
- [llvm] 804efba - PowerPC: Promote 32-bit ops to 64-bit in place for EXTSW elimination (#229054)
via llvm-commits
- [llvm] [Offload][omp] Use liboffload queues for AsyncInfo (PR #229713)
Alex Duran via llvm-commits
- [flang] [llvm] [flang] Fix wrong result for exponentiation when denormal is expected (PR #205262)
via llvm-commits
- [llvm] [AArch64][RegAlloc] Enable callee-saved register optimization for AArch64 (PR #214773)
Shreeyash Pandey via llvm-commits
- [llvm] [LV] Support fcmp in uncountable exit loops with stores (PR #228451)
Graham Hunter via llvm-commits
- [llvm] 1ebfb7a - [RISCV] Add Tenstorrent Ascalon XG (#229030)
via llvm-commits
- [llvm] [SCEV] Compare canonical SCEVs when grouping and folding operands. (PR #228224)
Florian Hahn via llvm-commits
- [clang] [llvm] [CMake] Honor symbol export lists for Emscripten (PR #229715)
Anutosh Bhat via llvm-commits
- [flang] [llvm] [AtomicExpand] Let targets keep the release fence out of the reservation (PR #214867)
Josef Schlehofer via llvm-commits
- [llvm] d6900ca - [Support] allow placement new with extended alignment (#228507)
via llvm-commits
- [llvm] 306576c - [TwoAddressInstruction] Don't mark a use undef if it reads a defined lane (#229554)
via llvm-commits
- [llvm] [Verifier] Check intrinsic signature validity before handwritten code (PR #229716)
Yingwei Zheng via llvm-commits
- [llvm] [Mips][FastISel] Also add low part when materializing private functions (PR #229717)
Alex Rønne Petersen via llvm-commits
- [clang] [llvm] [RISCV] Modernize tests for tt-ascalon-x (NFC) (PR #229718)
Ramkumar Ramachandra via llvm-commits
- [llvm] [mlir] [AArch64][RuntimeDyld] Handle far external symbols in COFF JIT code (PR #229719)
via llvm-commits
- [llvm] 01c1d2c - [ScheduleDAG](NFC) Allocate worklist once in `ScheduleDAGTopologicalSort` (#179256)
via llvm-commits
- [llvm] [ARM] don't warn when clobbering dregs (PR #229157)
Simon Tatham via llvm-commits
- [llvm] [offload][omp] Use olMemRegister for memory locking (PR #226450)
Alex Duran via llvm-commits
- [llvm] [SPIR-V] Type the format string of unmangled printf calls as i8 (PR #229720)
Tim Besard via llvm-commits
- [llvm] [SDAG] Reuse MMO in `legalizeStoreOps` promotion path (PR #229722)
Pierre van Houtryve via llvm-commits
- [llvm] AMDGPU/GlobalISel: Fix selecting i16 sext atomic loads on gfx6/gfx7 (PR #229723)
Matt Arsenault via llvm-commits
- [llvm] db9b959 - RuntimeLibcalls: Dedup dispatch calls to the same library function (#229561)
via llvm-commits
- [llvm] RuntimeLibcallsEmitter: Attribute DefaultCC to the library that references it (PR #229568)
Matt Arsenault via llvm-commits
- [llvm] RuntimeLibcalls: Add provider libraries for targets to reference (PR #229562)
Matt Arsenault via llvm-commits
- [llvm] [LLVM][CodeGen][SVE] Improve lowering for v1f32/f64 when NEON is not available. (PR #229724)
Paul Walker via llvm-commits
- [llvm] [CorrelatedValuePropagation] Drop stale debug records on context-only rewrites (PR #218498)
via llvm-commits
- [llvm] 2c88996 - [SPIR-V][NewPM] Port SPIRVModuleAnalysis (#227376)
via llvm-commits
- [llvm] [KnownFPClass] Add neg_square and fma_neg_square (PR #227248)
via llvm-commits
- [llvm] 9b989cb - [examples] Set target triple on modules in Kaleidoscope. (#229666)
via llvm-commits
- [llvm] [SCEV] Rework IncrementWrapFlags (PR #219559)
Ramkumar Ramachandra via llvm-commits
- [llvm] [LAA]Speculate byte strides to the access size (PR #227394)
Alexey Bataev via llvm-commits
- [llvm] 440cb1b - [LV] Support fcmp in uncountable exit loops with stores (#228451)
via llvm-commits
- [llvm] [MISched] Implement `UnanalyzableFrontier`-based DAG construction algorithm (PR #227375)
Nathan Corbyn via llvm-commits
- [clang] [llvm] [Clang] Update Codepoint names to Unicode 18. (PR #229080)
Corentin Jabot via llvm-commits
- [compiler-rt] [compiler-rt][ROCm] Collect Linux profiles from resident images (PR #229518)
Yaxun Liu via llvm-commits
- [clang] [llvm] [clang][Modules][Serialization] Avoid serializing duplicate source location entries for shared input files (PR #209795)
Ayokunle Amodu via llvm-commits
- [llvm] [Bazel] Fixes abbf695 (PR #229726)
via llvm-commits
- [llvm] [AMDGPU] Guard named barrier intrinsics with gfx12-insts (PR #227607)
Vikram Hegde via llvm-commits
- [llvm] [AMDGPU][NFC] Explicitly narrow conversions in AMDGPUBaseInfo (PR #215192)
via llvm-commits
- [llvm] [AMDGPU] Fix memory effects of tensor load/store intrinsics (PR #229728)
Yaxun Liu via llvm-commits
- [libc] [llvm] [libc] Add bare-metal Hexagon port (PR #217528)
via llvm-commits
- [llvm] [X86] Prefer MOVD/MOVSS over EXTRACTPS $0 for element 0 of v4f32 (PR #229730)
via llvm-commits
- [llvm] [LAA] Simplify pointer partitioning and speed up runtime-check grouping (PR #223758)
via llvm-commits
- [llvm] c1e0a94 - AMDGPU/GlobalISel: Fix selecting i16 sext atomic loads on gfx6/gfx7 (#229723)
via llvm-commits
- [clang-tools-extra] [flang] [libc] [llvm] [openmp] [docs] Repair remaining absolute self-documentation links (PR #228625)
Baranov Victor via llvm-commits
- [clang] [llvm] [HLSL] Add HLSL pack and pack_clamp intrinsics (PR #228089)
Alexander Johnston via llvm-commits
- [llvm] [libsycl] Fix prefetch pointer argument constness (PR #228442)
Sergey Semenov via llvm-commits
- [llvm] [MISched] Make SUnit operator<< debug only(NFC) (PR #229734)
Tomer Shafir via llvm-commits
- [llvm] [LowerMemIntrinsics] Expand small variable-length memsets inline (PR #228420)
Nathan Corbyn via llvm-commits
- [llvm] [LV] Rework collectUnitStridePredicates (PR #216362)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SPIR-V] Move MIR rewrites out of SPIRVModuleAnalysis (PR #229736)
Arseniy Obolenskiy via llvm-commits
- [llvm] [CodeGen][AMDGPU][NPM] Wire up SSAMachineSchedulerPass in optimized regalloc pipeline (PR #229737)
Vikram Hegde via llvm-commits
- [llvm] [AMDGPU] Don't select scalar loads for LDS/scratch (PR #227423)
Matt Arsenault via llvm-commits
- [llvm] [gn build] Port 2c10250e85242 (PR #229744)
Nico Weber via llvm-commits
- [llvm] [gn build] Port e03123a631628 (PR #229745)
Nico Weber via llvm-commits
- [llvm] [gn build] Port e128a79c81bd4 (PR #229746)
Nico Weber via llvm-commits
- [llvm] [gn] Add include_dir required by abbf6958bb8449 (PR #229747)
Nico Weber via llvm-commits
- [llvm] [BOLT][AArch64] Add support for conditional tail calls in cold code (PR #227693)
Alexandros Lamprineas via llvm-commits
- [llvm] [LAA] Avoid stray predicates from replaceSymStrides (PR #216350)
Florian Hahn via llvm-commits
- [llvm] 7972124 - [gn] Add include_dir required by abbf6958bb8449 (#229747)
via llvm-commits
- [llvm] [PowerPC] Add regression test for #229620 (PR #229748)
Aochang Liu via llvm-commits
- [llvm] 39e103c - [gn build] Port e128a79c81bd4 (#229746)
via llvm-commits
- [llvm] PPC: Fix unnecessary host requirement in test (PR #229749)
Matt Arsenault via llvm-commits
- [llvm] Wasm nfc tests (PR #229750)
Sam Parker via llvm-commits
- [llvm] PPC: Fix EXTSW elimination promoting a subregister operand (PR #208051)
Matt Arsenault via llvm-commits
- [llvm] [offload][omp] Remove data_fence interfaces (PR #226445)
Alex Duran via llvm-commits
- [llvm] [AMDGPU][GlobalISel] Legalize bf16 for fadd, fmul, fma, fsub, fmin, fmax, and fcanonicalize (PR #227901)
via llvm-commits
- [llvm] 9ee0c90 - [AArch64] Avoid materializing full masks for extractelement users of predicate-as-counter (#220961)
via llvm-commits
- [llvm] [MachineLateInstrsCleanup] Reuse redundant spill-slot reloads (PR #220180)
Anup Kumar via llvm-commits
- [llvm] [AArch64] Use multi-vector intrinsics for masked load/store users of predicate-as-counter (PR #220962)
Benjamin Maxwell via llvm-commits
- [llvm] [MLGO] Avoid NaN regalloc features for unspillable live ranges (PR #229751)
Arseniy Obolenskiy via llvm-commits
- [llvm] [Bazel] Add libraries for bug reporting plugins (not in list yet) (PR #229474)
David Young via llvm-commits
- [llvm] [KnownFPClass] Teach KnownFPClass that tanh(non-positive) = non-positive (PR #229752)
via llvm-commits
- [llvm] 7ca0478 - [Verifier] Check intrinsic signature validity before handwritten code (#229716)
via llvm-commits
- [llvm] [LAA] Use diff checks for read-modify-write pointers (PR #226816)
Dávid Bolvanský via llvm-commits
- [llvm] [LoopFusion] Initial Cost Model (PR #224626)
via llvm-commits
- [llvm] [GlobalISel] Add G_MUL to computeNumSignBits (PR #229766)
Tim Gymnich via llvm-commits
- [llvm] [ValueTracking][InstCombine] Fold add/or/xor to disjoint or using assume(no common bits) (PR #223344)
via llvm-commits
- [clang] [llvm] [IRBuilder] Use getDataLayout() in more places (NFC) (PR #229767)
Nikita Popov via llvm-commits
- [llvm] [CodeGen] Fix physical-register input latency for inline asm (PR #229764)
Alan Li via llvm-commits
- [llvm] cd6eae7 - [AMDGPU][GlobalISel] Legalize bf16 for fadd, fmul, fma, fsub, fmin, fmax, and fcanonicalize (#227901)
via llvm-commits
- [llvm] 1d3c9f0 - [GlobalISel] [AArch64] Generate MVN for shift by (N - X) in SelectShiftMask (#226273)
via llvm-commits
- [llvm] [OptBisect] Add function filtering support to opt-bisect (PR #222123)
Daniel Donenfeld via llvm-commits
- [llvm] 1b17950 - [IRBuilder] Use getDataLayout() in more places (NFC) (#229767)
via llvm-commits
- [llvm] [X86] Use F16C for contract fptrunc from double to half (PR #229782)
Aochang Liu via llvm-commits
- [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Connor Tsui via llvm-commits
- [llvm] [AArch64] SVE Shuffleopt: merge reduction reverse into tbl (PR #206047)
Graham Hunter via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
Paul Walker via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
Paul Walker via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
Paul Walker via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
Paul Walker via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [AArch64] Share materialized constant between add x, k and sub x,k (PR #225430)
via llvm-commits
- [llvm] [NewPM] Sync passes names between legacy and NewPM (PR #229775)
Sam Elliott via llvm-commits
- [llvm] 22d7058 - [GlobalISel] Add G_MUL to computeNumSignBits (#229766)
via llvm-commits
- [clang] [llvm] [HLSL] Add handling of Interpolation Modifiers (PR #225452)
Finn Plummer via llvm-commits
- [llvm] [LV] Adjust the trip count of tail-folded loops which may overflow (PR #224340)
Kerry McLaughlin via llvm-commits
- [llvm] [AMDGPU][NFC] Precommit tests for WMMA new hazard CFG walk (PR #229790)
Janek van Oirschot via llvm-commits
- [llvm] [MC] Make MCAsmStreamer::addExplicitComment agnostic to line-endings (PR #229791)
via llvm-commits
- [clang] [llvm] [clang][NVPTX] Emit !atomic.ignore.denormal.mode for CUDA atomics (PR #217587)
Christian Sigg via llvm-commits
- [clang] [llvm] [HLSL] Add semantic signature packing flags and invoke packing (PR #226554)
Finn Plummer via llvm-commits
- [llvm] [AMDGPU] Allow merging v_mov_bf16 across mid-block exec writes (aka WWM spills) (PR #229265)
Jay Foad via llvm-commits
- [llvm] [SPIRV] Preserve loop hints for loops with multiple exits (PR #229590)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU] Fold high-half 16-bit extracts into VOP3 op_sel (PR #222255)
Matt Arsenault via llvm-commits
- [llvm] [AArch64][GlobalISel] Remove more scalar types from tests. NFC (PR #229807)
Tim Gymnich via llvm-commits
- [clang] [llvm] [Clang][AArch64] Command-line options for A-profile's Return Address Authentication Hardening (PR #176171)
Victor Campos via llvm-commits
- [llvm] [DirectX] Move the DXContainer debug options from MC to DirectX (PR #228329)
Farzon Lotfi via llvm-commits
- [llvm] CodeGen: Run LiveIntervals before PHIElimination and drop LiveVariables from it (PR #228618)
Matt Arsenault via llvm-commits
- [flang] [llvm] [mlir] [flang][mlir][OpenMP] Report Fortran names for privatized target maps… (PR #229823)
Jason Van Beusekom via llvm-commits
- [llvm] AMDGPU: Fix legacy triples that have appeared in new tests (PR #229837)
Matt Arsenault via llvm-commits
- [llvm] [Bazel] Fixes 636bedc (PR #229838)
via llvm-commits
- [llvm] [SeparateConstOffsetFromGEP] Track cast state during offset extraction (PR #229839)
Alex MacLean via llvm-commits
- [llvm] [X86][NewPM] Pass correct class to X86FixupVectorConstantsPass (PR #229840)
Mikołaj Piróg via llvm-commits
- [llvm] [LoongArch] Enable callee saved register optimization (PR #226809)
via llvm-commits
- [llvm] [SLP] Merge redundant input operands of binary-operator nodes (PR #208670)
Balakrishna Bandlapalli via llvm-commits
- [llvm] [InstCombine] Fold fdiv by splat of pow/exp/powi into fmul (PR #227238)
Harrison Hao via llvm-commits
- [llvm] [AMDGPU] Remove some dead code (NFC) (PR #229821)
Arseniy Obolenskiy via llvm-commits
- [llvm] 0e49a2c - [SPIRV] Preserve loop hints for loops with multiple exits (#229590)
via llvm-commits
- [llvm] [AArch64][TTI] Tune costs for mulh. (PR #225788)
David Green via llvm-commits
- [llvm] [GVN] Do not create/destroy a local copy of GVNPassImpl in the run method (PR #229833)
Momchil Velikov via llvm-commits
- [llvm] [InstCombine] Use UTC for ffs-1.ll (PR #229841)
Aiden Grossman via llvm-commits
- [llvm] [AArch64] Better bitcast costs. (PR #227354)
David Green via llvm-commits
- [llvm] [NVPTX] Use ExpandIRInsts for frem (PR #224141)
Alex MacLean via llvm-commits
- [llvm] [DAGCombiner] Widen zero-extended low-bits masks to select and-not (PR #228910)
Dávid Bolvanský via llvm-commits
- [llvm] [ProfCheck] Remove already fixed tests (PR #229819)
Aiden Grossman via llvm-commits
- [clang] [llvm] [clang][OpenMP] Fix assertion in vtable registration for incomplete types (PR #229521)
David Pagan via llvm-commits
- [llvm] [LV] Handle complex multiply reductions (PR #204349)
via llvm-commits
- [llvm] [SLP] NFC: make candidate state ownership explicit (PR #229846)
Alexey Bataev via llvm-commits
- [clang] [flang] [llvm] [CodeGen] Honor -function-splitting=none in BasicBlockSections (PR #227880)
Rahman Lavaee via llvm-commits
- [llvm] [AArch64][GlobalISel] Fix LLT types in test (PR #229866)
Tim Gymnich via llvm-commits
- [llvm] [RISCV] Replace hasStdExtC() with hasStdExtZcf() && hasStdExtZcd(). (PR #229870)
Craig Topper via llvm-commits
- [llvm] AMDGPU: Insert readfirstlane for write_register of a VGPR value (PR #229854)
LLVM Continuous Integration via llvm-commits
- [llvm] [AMDGPU] Expand constant named barrier matching in SDag (PR #191506)
Krzysztof Drewniak via llvm-commits
- [llvm] 70f06c9 - [CMake] Honor symbol export lists for Emscripten (#229715)
via llvm-commits
- [llvm] [Local] Construct IRBuilder in ConstantFoldTerm only when folding (NFC) (PR #229437)
Nikita Popov via llvm-commits
- [lld] [llvm] [lld][llvm-objcopy] Fix iterator invalidation when mutating vectors during iteration (PR #228183)
via llvm-commits
- [llvm] de374c6 - [RISCV] Add mask version of vunzipe/o.v (#228597)
via llvm-commits
- [clang] [lldb] [llvm] [OpenCL][Clang] Add support for cooperative matrix extension (PR #221328)
via llvm-commits
- [llvm] aff82f1 - [LV] Don't vectorize loops using swifterror values. (#229875)
via llvm-commits
- [llvm] [LV] Don't vectorize loops using swifterror values. (PR #229875)
Florian Hahn via llvm-commits
- [llvm] [SPIRV] Preserve AMDGPU metadata through NonSemantic.AuxData (PR #215510)
Marcos Maronas via llvm-commits
- [llvm] Cvt scale analysis (PR #229889)
Justin Rosner via llvm-commits
- [llvm] [RISCV] Don't infer C or Zce in updateCZceFeatureImplications. (PR #229905)
Craig Topper via llvm-commits
- [llvm] [AMDGPU] Allow commuting immediates out of src0 when legal (PR #229906)
Adel Ejjeh via llvm-commits
- [llvm] [SelectionDAG] Do not treat variadic vreg results as physreg defs (PR #229847)
Derek Schuff via llvm-commits
- [llvm] 55c58be - [SimplifyLibCalls] Fix profiles in optimizeFFS (#229842)
via llvm-commits
- [llvm] [SimplifyLibCalls] Fix profiles in optimizeFFS (PR #229842)
Aiden Grossman via llvm-commits
- [llvm] [InstCombine] Fix profile propagation in ldexp.ll (PR #229908)
Aiden Grossman via llvm-commits
- [llvm] [LAA] Add baseline tests and regenerate checks for RMW diff checks (PR #229774)
Ramkumar Ramachandra via llvm-commits
- [llvm] [ARM] Replace reg in preheader when reverting WhileLoopStartLR (PR #229754)
David Green via llvm-commits
- [llvm] [KnownFPClass] Fix `KnownFPClass::log` deductions for ppcf128 (PR #223296)
Eric Ross via llvm-commits
- [llvm] [MemProf] Match call sites with line 0 (C++ non-virtual thunks) (PR #229919)
David Li via llvm-commits
- [llvm] [InstCombine] Generalize sext (trunc (lshr Y, C)) -> sext/trunc (ashr Y, C) to other shift amounts (PR #227320)
Luke Lau via llvm-commits
- [llvm] [AMDGPU] Add a static performance simulator for GFX1250 (PR #229922)
via llvm-commits
- [llvm] WebAssembly: Take the exception model from the module flag (PR #227583)
Matt Arsenault via llvm-commits
- [llvm] 66a38a5 - [RISCV] Ensure Shadow Stack checks correct RA (#226309)
via llvm-commits
- [llvm] [VPlan] Support tailfolded loops in multi-use-reductions (PR #214455)
Florian Hahn via llvm-commits
- [llvm] AMDGPU: Stop relying on -amdgpu-scalarize-global-loads=false in ALU tests (PR #229924)
Matt Arsenault via llvm-commits
- [llvm] 81f7e67 - [LV] Add tests showing alloca miscompiles. (#229911)
via llvm-commits
- [llvm] [LV] Add tests showing alloca miscompiles. (PR #229911)
Florian Hahn via llvm-commits
- [llvm] [VPlan] Re-run makeScalarizationDecisions after call widening. (PR #229925)
Florian Hahn via llvm-commits
- [llvm] [KnownFPClass] Combine fpext and fptrunc into a shared routine (PR #225188)
Eric Ross via llvm-commits
- [llvm] [AMDGPU] Analyze subword loads in redundant AND deletion (PR #229926)
via llvm-commits
- [llvm] [LV] Have ExtractValue match the shape of the operand. (PR #229927)
Florian Hahn via llvm-commits
- [lld] [lld][MachO][NFC] Prepare ObjC stubs for variable sizes (PR #219742)
Ellis Hoag via llvm-commits
- [llvm] [Plugins] Always export `llvmGetPassPluginInfo()` (PR #229801)
Jared Wyles via llvm-commits
- [clang] [llvm] [InstCombine] Relax singleUse for GEP(p,add(x,constant)) to 2 GEPs (PR #209551)
Abhay Kanhere via llvm-commits
- [clang] [llvm] [CodeView] Set the S_COMPILE3 hotpatch flag from a module flag (PR #229868)
Daniel Paoliello via llvm-commits
- [llvm] [Offload][AMDGPU] Wire HSA profiling into GenericProfiler abstraction (PR #226044)
Jan Patrick Lehr via llvm-commits
- [llvm] 33ec29c - [MLGO][Docs] Document MLIR based model embedding (#227536)
via llvm-commits
- [clang] [flang] [llvm] [flang] Diagnose and optionally repair missing MODULE procedure prefixes (PR #220783)
Andre Kuhlenschmidt via llvm-commits
- [llvm] 07536a4 - [InstCombine] Generalize sext (trunc (lshr Y, C)) -> sext/trunc (ashr Y, C) to other shift amounts (#227320)
via llvm-commits
- [llvm] 9594c0d - [InstCombine] Move sext(trunc(...)) combines before EvaluateInDifferentType (#227321)
via llvm-commits
- [llvm] 52c1143 - [InstCombine] Move sext(trunc(lshr(...))) combine before same type combine (#227322)
via llvm-commits
- [llvm] [TableGen] Fix a strict weak ordering bug (PR #229955)
Alan Zhao via llvm-commits
- [llvm] [flang-rt] Fixed usage of strtok_r. (PR #229570)
Slava Zakharin via llvm-commits
- [llvm] [AMDGPU]Select build_vector to v_pk_mov (PR #221673)
via llvm-commits
- [llvm] [DAG][AMDGPU] Add folding ISD::SELECT from vXiY into vZi32 with X * Y = Z * 32 (PR #173328)
via llvm-commits
- [llvm] [CodeGen] Drop dead SlotIndexes before allocation (PR #224031)
Petar Jovanovic via llvm-commits
- [llvm] [AMDGPU] Adjust amdgpu fmax/fmin legalization for GISEL (PR #223351)
via llvm-commits
- [llvm] [orc-rt] Add SymbolLookupFlags and SymbolLookupSet utilities (PR #229958)
Lang Hames via llvm-commits
- [llvm] [LV] Don't narrow Widen-IV with wider use with scalar VF. (PR #229960)
Elvis Wang via llvm-commits
- [llvm] [DirectX] Lower semantic signature metadata to DXContainer parts and DXIL metadata (PR #226562)
Deric C. via llvm-commits
- [llvm] [InstCombine] Mark profiles unknown when folding lshr shifts (PR #229965)
Aiden Grossman via llvm-commits
- [llvm] [SLP] Use costing to decide between Strided Loads and Compressed Loads (PR #226540)
Ryan Buchner via llvm-commits
- [llvm] dc488b8 - [Clang][AArch64] Command-line options for A-profile's Return Address Authentication Hardening (#176171)
via llvm-commits
- [llvm] [ISel] Add extension attribute on Memset libcall arg. (PR #229883)
Jonas Paulsson via llvm-commits
- [llvm] [flang-rt] Fix MOD/MODULO for REAL operands of magnitude 2**63 (PR #229895)
Eugene Epshteyn via llvm-commits
- [llvm] [Mips] Track ISA mode and distinguish ELF data labels (PR #228883)
Jiaxun Yang via llvm-commits
- [llvm] [OpenMP][OMPTarget] Avoid Shadow Pointer post-processing for FROM and UPDATE TO unless required (PR #229966)
via llvm-commits
- [llvm] 85b4fbf - [Mips][FastISel] Also add low part when materializing private functions (#229717)
via llvm-commits
- [llvm] [Mips] Lower calls to private functions the same way as internal functions (PR #229975)
Alex Rønne Petersen via llvm-commits
- [llvm] e9b9fca - [Mips] Track ISA mode and distinguish ELF data labels (#228883)
via llvm-commits
- [clang] [llvm] [clang-cl] Use MSVC's "GS buffer" heuristic for the default /GS (PR #225626)
Pan Tao via llvm-commits
- [lld] [ELF] Simplify Eq/EEq options with SeparateOrEq. NFC (PR #229977)
Fangrui Song via llvm-commits
- [llvm] [AArch64] Ignore processor-family tuning features for inlining (PR #229979)
Chris Kennelly via llvm-commits
- [llvm] [msan] Propagate shadow through addrspacecast. (PR #229860)
Vitaly Buka via llvm-commits
- [llvm] [SCEV] Clear AddRec no-wrap flags when forgetting results (PR #212744)
via llvm-commits
- [llvm] [InstCombine] Drop wrong ninf for `(-X / Y) + Z` (PR #228684)
via llvm-commits
- [llvm] [AMDGPU] Ignore debug instructions in invertSCCUse scan budget (PR #229961)
Yongqiang Tian via llvm-commits
- [llvm] [DAGCombiner] Add tests for zero-extended low-bits masks (NFC) (PR #229981)
Chris Kennelly via llvm-commits
- [llvm] [KnownFPClass] Add KnownFPClass functions for trunc, floor, and ceil (PR #229982)
Eric Ross via llvm-commits
- [llvm] 3ed4eae - [OpenMPOpt][PGO] Set entry count on merged parallel wrapper (#221654)
via llvm-commits
- [llvm] [OpenMPOpt][PGO] Set entry count on merged parallel wrapper (PR #221654)
Alok Kumar Sharma via llvm-commits
- [llvm] 6d86521 - [X86] Make EXTRQI/INSERTQI SDTypeProfile accurate and enable verifyTargetNode (#229655)
via llvm-commits
- [llvm] fd35b54 - [RISCV] Remove unnecessary redeclarations of XLenVT in LowerINTRINSIC_WO_CHAIN. NFC (#228152)
via llvm-commits
- [llvm] [RISCV] Remove unnecessary redeclarations of XLenVT in LowerINTRINSIC_WO_CHAIN. NFC (PR #228152)
Craig Topper via llvm-commits
- [llvm] [Offload] Unlock mapping entry before deleting it in disassociatePtr (PR #223319)
via llvm-commits
- [compiler-rt] [sanitizer] Always report invalid aligned operator new requests (PR #225847)
Vitaly Buka via llvm-commits
- [llvm] RISCV: Do not add a live VL def to inline asm that clobbers it (PR #229246)
Pengcheng Wang via llvm-commits
- [llvm] [PowerPC] Accept .set directive in mem operand (PR #229992)
via llvm-commits
- [llvm] [ExpandReductions] Add support for non-pow2 vector counts (PR #208738)
Prajwal KP via llvm-commits
- [llvm] [AMDGPU][DAG] Allow extract_subvector for free register copy (PR #214152)
via llvm-commits
- [llvm] [AArch64] Ignore debug instructions in MachineOutliner range partitioning (PR #229993)
Yongqiang Tian via llvm-commits
- [llvm] [PowerPC] Ignore debug uses when removing dead CRSET/CRUNSET (PR #229994)
Yongqiang Tian via llvm-commits
- [llvm] [NFC][CodeGen] Replace memcpy with range-based copies (PR #229461)
Younes Yazami via llvm-commits
- [llvm] [Offload] Check Expected from hasPendingWorkImpl in AMDGPU dataFill (PR #223329)
via llvm-commits
- [llvm] [Offload] Return an integer user count for AMDGPU HSA queues (PR #223335)
via llvm-commits
- [llvm] [Offload] Check HSA allocate status before using the pointer (PR #223337)
via llvm-commits
- [llvm] [Offload] Avoid writing past the destination in host dataFill (PR #223331)
via llvm-commits
- [llvm] [RISCV] Only expand lla to qc.e.li in .option nopic (PR #229995)
Sam Elliott via llvm-commits
- [llvm] [CodeGen] Force compiler to generate default destructor for class (PR #229944)
Fabrice de Gans via llvm-commits
- [llvm] 9f20fc9 - [Plugins] Always export `llvmGetPassPluginInfo()` (#229801)
via llvm-commits
- [llvm] [X86][LoopVectorize] Enable MaximizeBandwidth by default on X86 (PR #229997)
Sumukh J Bharadwaj via llvm-commits
- [llvm] 34654fc - [AMDGPU] Ignore debug instructions in invertSCCUse scan budget (#229961)
via llvm-commits
- [llvm] [RISCV] Fold bounded e64 vector adds into widening forms (PR #229831)
via llvm-commits
- [llvm] [RISCV] Model Zvzip costs for fixed-length shuffles (PR #227677)
via llvm-commits
- [llvm] c279be8 - RISCV: Do not add a live VL def to inline asm that clobbers it (#229246)
via llvm-commits
- [llvm] [Target][NFC] Use SmallPtrSet instead of SmallSet for pointers (PR #230002)
Ben Wu via llvm-commits
- [llvm] [RISCV] Change custom promotion of riscv_clmulh/r (PR #228846)
Pengcheng Wang via llvm-commits
- [llvm] [RISCV] Combine (srl (clmul (and X, 0xffffffff), (and Y, 0xffffffff)), 31). (PR #228848)
Pengcheng Wang via llvm-commits
- [llvm] 0ae305d - AArch64: Test trap-after-noreturn against the exception model (#229681)
via llvm-commits
- [llvm] [IndVars] Rewrite loop-exit comparisons from operands (PR #223951)
Pengcheng Wang via llvm-commits
- [llvm] [DebugInfo][Reassociate] Drop stale locations from moved instructions (PR #230005)
via llvm-commits
- [llvm] [AMDGPU] Let SIOptimizeVGPRLiveRange handle AV-class virtual registers (PR #226527)
via llvm-commits
- [llvm] [DAGCombiner] Defer store merging (PR #224879)
Pengcheng Wang via llvm-commits
- [llvm] 4264f65 - X86: Use hasOneNonDBGUse instead of kill flag for cascaded select (#228939)
via llvm-commits
- [llvm] X86: Use hasOneNonDBGUse instead of kill flag for cascaded select (PR #228939)
Matt Arsenault via llvm-commits
- [llvm] [LAA] Rework getStrideFromPointer, improve speculation (PR #212816)
Ramkumar Ramachandra via llvm-commits
- [llvm] AMDGPU: Migrate more tests off of -amdgpu-scalarize-global-loads=false (PR #230008)
Matt Arsenault via llvm-commits
- [llvm] [AArch64] Allow single-element vector FP converts with +fprcvt (PR #169692)
via llvm-commits
- [llvm] [CodeGen] Remove unused IntrinsicLowering::LowerToByteSwap (NFC) (PR #230010)
Kazu Hirata via llvm-commits
- [llvm] 2d2fbad - [SPIR-V] Type the format string of unmangled printf calls as i8 (#229720)
via llvm-commits
- [clang] [llvm] [BasicAA] Fix miscompilation with setjmp/longjmp due to missing longjmp re-entry paths in alias analysis (PR #212297)
via llvm-commits
- [llvm] [RegAllocFast] Add implicit super-register operands after operand loops. NFC (PR #230011)
Fangrui Song via llvm-commits
- [llvm] e684d6d - [AArch64] Fix register offset ld/st predicate (#225645)
via llvm-commits
- [llvm] [AArch64] Fix register offset ld/st predicate (PR #225645)
Tomer Shafir via llvm-commits
- [llvm] [NFC][AMDGPU] Add tests for the cost of vector f32 to f16 fptrunc (PR #229949)
via llvm-commits
- [llvm] 1a79788 - [X86] Remove x86-widen-shift flag (#229353)
via llvm-commits
- [llvm] [LAA] Strip Assume API of getPtrStride (NFC) (PR #229836)
Nikita Popov via llvm-commits
- [llvm] d23ae3d - AMDGPU: Remove promotion of 32 and 64-bit atomic load/store types (#229696)
via llvm-commits
- [llvm] AMDGPU: Remove promotion of 32 and 64-bit atomic load/store types (PR #229696)
Matt Arsenault via llvm-commits
- [llvm] [AArch64][GlobalISel] More test coverage for vector immediates. NFC (PR #229928)
David Green via llvm-commits
- [llvm] AMDGPU: Avoid null and poison pointers in call-argument-types (PR #230012)
Matt Arsenault via llvm-commits
- [llvm] [RISCV] Avoid early split for fixed-length VECTOR_DEINTERLEAVE (PR #226453)
via llvm-commits
- [llvm] [AMDGPU][GlobalISel] Assign bank to vregs created by RegBankCombiner (PR #216270)
Arseniy Obolenskiy via llvm-commits
- [llvm] [SimplifyCFG] Widen `!range` over PassThru in hoistConditionalLoadsStores (PR #217821)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU] Require true low byte extensions in dot4 mul operands (PR #220169)
Arseniy Obolenskiy via llvm-commits
- [llvm] AMDGPU: Stop relying on -amdgpu-scalarize-global-loads=false in more tests (PR #230013)
Matt Arsenault via llvm-commits
- [llvm] [RISCV] Declare command line options in TableGen (PR #230014)
Pengcheng Wang via llvm-commits
- [llvm] [mlir] [polly] [IR] Deprecate Instruction::getStableDebugLoc() (PR #230016)
Antonio Frighetto via llvm-commits
- [llvm] [VPlan] Don't merge allocas across lanes, parts or recipes. (PR #229929)
David Sherwood via llvm-commits
- [llvm] 1514c1f - [NFC][AutoUpgrade] Add logging for module flags upgrade (#226165)
via llvm-commits
- [llvm] [NFC][AutoUpgrade] Add logging for module flags upgrade (PR #226165)
Andrew Savonichev via llvm-commits
- [llvm] [SPIR-V] Remove some dead code (NFC) (PR #230019)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AArch64][SVE] Fix predicate type for quadword gather/scatter SDNodes (PR #228274)
Sander de Smalen via llvm-commits
- [llvm] [Passes] Declare command line options in TableGen (PR #230020)
Fangrui Song via llvm-commits
- [llvm] 1946745 - [CodeGen] Remove unused IntrinsicLowering::LowerToByteSwap (NFC) (#230010)
via llvm-commits
- [llvm] [AMDGPU][GFX13] Add codegen for indexed VBUFFER resources (PR #230015)
via llvm-commits
- [llvm] [NFC][MC] Replace memcpy with range-based copies (PR #229466)
James Henderson via llvm-commits
- [llvm] [TableGen][AMDGPU] Introduce register sequences (PR #226106)
Ivan Kosarev via llvm-commits
- [llvm] [TableGen] Share identical super-register class masks (PR #224633)
Ivan Kosarev via llvm-commits
- [llvm] [VectorUtils] Pad concatenateVectors with poison instead of undef (PR #225598)
Nikita Popov via llvm-commits
- [llvm] 4810256 - [PGO] Fix PGO instrumentation for selects with poison condition (#229002)
via llvm-commits
- [llvm] [BOLT][RISCV] Ignore alignment markers in linked binaries (PR #223994)
via llvm-commits
- [llvm] [BOLT][RISCV] Emit PCREL_LO12 labels during size estimation (PR #224202)
via llvm-commits
- [llvm] [CodeGen] Don't rotate shared increment blocks to loop top (PR #219126)
Matt Arsenault via llvm-commits
- [llvm] 8b423ff - AMDGPU: Avoid null and poison pointers in call-argument-types (#230012)
via llvm-commits
- [llvm] [Object] Move the createBinary/createObjectFile factories into their own file (PR #229795)
James Henderson via llvm-commits
- [llvm] github-upload-release: Update script to work with draft releases (PR #229909)
Cullen Rhodes via llvm-commits
- [llvm] 7733db3 - AMDGPU: Stop relying on -amdgpu-scalarize-global-loads=false in more tests (#230013)
via llvm-commits
- [llvm] cb83974 - [LAA] Strip Assume API of getPtrStride (NFC) (#229836)
via llvm-commits
- [llvm] MC: Error if target did not register null streamer (PR #230024)
Matt Arsenault via llvm-commits
- [llvm] [RISCV][GlobalISel] Promote f16 to f32 for G_FPTOSI/G_FPTOUI without Zfh (PR #230025)
Kane Wang via llvm-commits
- [flang] [llvm] [flang] Use backend fp128 support to determine REAL(16) availability (PR #221907)
Shunsuke Watanabe via llvm-commits
- [llvm] [AMDGPU] Fold add of a variable into a zero dot accumulator (PR #230026)
Harrison Hao via llvm-commits
- [llvm] [VPlan] Introduce WrapFlagsTy::withoutNSW (NFC) (PR #230030)
via llvm-commits
- [llvm] 62a04f9 - [IR] Deprecate Instruction::getStableDebugLoc() (#230016)
via llvm-commits
- [llvm] [BOLT][RISCV] Add RISCV tests for external branch target validation (PR #223991)
Jinjie Huang via llvm-commits
- [llvm] [RISCV] Use ReadJmp for XAndesPerf branch instructions (PR #230000)
Pengcheng Wang via llvm-commits
- [llvm] [SandboxIR] Move IRBuilder from Context to Module (NFC) (PR #230033)
Nikita Popov via llvm-commits
- [llvm] 7ac4051 - [VPlan] Add getLastOperand helper (NFC). (#229354)
via llvm-commits
- [llvm] [AMDGPU] Overload rsrc operand of non-ptr buffer intrinsics (PR #230034)
via llvm-commits
- [llvm] [CodeGen] Only visit live units in ::removeRegsNotPreserved (NFC) (PR #230035)
Florian Hahn via llvm-commits
- [llvm] RegisterCoalescer: Keep remat def dead if it's a copy destination superregister (PR #230036)
Matt Arsenault via llvm-commits
- [llvm] [X86] Add mayFoldIntoScalarInteger helper. NFC. (PR #230037)
Simon Pilgrim via llvm-commits
- [llvm] [X86] Form (FM)ADDSUB for a chain of alternating FSUB/FADD (PR #226429)
Tim Besard via llvm-commits
- [llvm] 45c4f0f - [GlobalISel][AArch64] Lower G_BITCAST vectors with non-integer element ratio (#215154)
via llvm-commits
- [llvm] [VPlan] Add test coverage for CSE-rdx (PR #229732)
Florian Hahn via llvm-commits
- [llvm] cbbae6e - [RISCV] Use ReadJmp for XAndesPerf branch instructions (#230000)
via llvm-commits
- [llvm] [RISCV] Preserve independent VTYPE facts across CFG joins (PR #228735)
Luke Lau via llvm-commits
- [llvm] [LV] Add baseline tests for predicated load access safety (PR #229468)
Florian Hahn via llvm-commits
- [llvm] [AArch64] Fold DUPLANE of 0 to zeroimm (PR #228415)
David Sherwood via llvm-commits
- [llvm] 91096d9 - [AArch64][GlobalISel] Ensure narrowed loads are given integer types. (#229857)
via llvm-commits
- [llvm] [AArch64][GlobalISel] Ensure narrowed loads are given integer types. (PR #229857)
David Green via llvm-commits
- [llvm] InjectTLIMappings: skip TLI mapping injection for variadic calls. (PR #202016)
David Sherwood via llvm-commits
- [llvm] 7cfe81b - [X86] Don't treat vectors with elements wider than 64 bits as immediate-shiftable (#228840)
via llvm-commits
- [llvm] 03bc769 - [AArch64][TTI] Tune costs for mulh. (#225788)
via llvm-commits
- [llvm] [AMDGPU] Promote allocas used by phi/select (PR #221870)
via llvm-commits
- [llvm] [1/7][PISA] Add PISA target registration and backend stub (PR #214094)
Michal Paszkowski via llvm-commits
- [llvm] 7451898 - [SPIR-V] Lower clang __to_global, __to_local and __to_private builtins (#229211)
via llvm-commits
- [llvm] CodeGen: Compute LiveIntervals before TwoAddressInstructions (PR #225174)
Doha Nam via llvm-commits
- [lld] [llvm] [lld][DebugInfo/BTF] Add BTF section merging and deduplication (PR #183915)
Can H. Tartanoglu via llvm-commits
- [llvm] [CodeGenPrepare] Split x/select(cond,y,C) to select(cond,x/y,x/C) (PR #215967)
via llvm-commits
- [llvm] [DAG] Use getLegalMaskAndStepVector in expandGetActiveLaneMask (PR #226506)
Luke Lau via llvm-commits
- [llvm] [X86] combineSelect - don't check SimpleValueTypes through bitcast (PR #230040)
Simon Pilgrim via llvm-commits
- [llvm] a6706e1 - [VPlan] Introduce WrapFlagsTy::withoutNSW (NFC) (#230030)
via llvm-commits
- [llvm] [VPlan] Verify canonical IV has no users if there's a VPCurrentIterationPHI. NFC (PR #197431)
Luke Lau via llvm-commits
- [llvm] 194bcb1 - [X86] Add mayFoldIntoScalarInteger helper. NFC. (#230037)
via llvm-commits
- [llvm] [cmake] Limit `/Zc:dllexportInlines-` flag to C/C++ (PR #227791)
Fabrice de Gans via llvm-commits
- [clang-tools-extra] [llvm] [cmake] Properly link against LLVM components (PR #224624)
Fabrice de Gans via llvm-commits
- [llvm] a891455 - [DAGCombiner] Add tests for zero-extended low-bits masks (NFC) (#229981)
via llvm-commits
- [llvm] BitcodeReader: check strtab bounds without overflow in partition handling (PR #230048)
via llvm-commits
- [llvm] dcb4cd7 - [LV] Add baseline tests for predicated load access safety (#229468)
via llvm-commits
- [llvm] [SCEV] Look for guard conditions through merge blocks. (PR #228135)
Florian Hahn via llvm-commits
- [llvm] [BOLT][AArch64] Error when LongJmp will relax beyond 128MiB (PR #229882)
Adam Bzowski via llvm-commits
- [llvm] [X86] Compute `CTLZ` using fp math for `v4i32` on SSE2 targets (PR #190930)
Simon Pilgrim via llvm-commits
- [llvm] [InstCombine] Remove target specific sext(trunc(...)) combine tests. NFC (PR #230049)
Luke Lau via llvm-commits
- [llvm] 5f890fc - [PassManager] Factor out some common code to a utility function (NFC) (#229432)
via llvm-commits
- [llvm] [PassManager] Factor out some common code to a utility function (NFC) (PR #229432)
Momchil Velikov via llvm-commits
- [compiler-rt] [sanitizer_common] Add unit tests for operator new failure handling (PR #226832)
Justin T. Gibbs via llvm-commits
- [llvm] [X86] Fold extensions into atomic loads (PR #217886)
Simon Pilgrim via llvm-commits
- [llvm] [AArch64][CostModel] Add cost model for pdep/pext instructions (PR #207657)
Yashwant Singh via llvm-commits
- [llvm] c85b25d - [AArch64] Implement CRC32 const folding (#228985)
via llvm-commits
- [llvm] [LV][CSSPGO] Don't let pseudo probes block early-exit vectorization (PR #220149)
Luke Lau via llvm-commits
- [llvm] [IR][RFC] Add @llvm.mask.beforefirst intrinsic (PR #203874)
Luke Lau via llvm-commits
- [compiler-rt] [llvm] [asan] Adopt sanitizer_common operator-new framework (PR #196388)
Justin T. Gibbs via llvm-commits
- [llvm] 93c9c6b - [NVVM][NVPTX] Support spcompress and spdecompress intrinsics (#221751)
via llvm-commits
- [llvm] [WebAssembly] Lower v4i8 mul (PR #230052)
Sam Parker via llvm-commits
- [llvm] [X86] SimplifyDemandedBitsForTargetNode - add X86ISD::BZHI handling (PR #199828)
Simon Pilgrim via llvm-commits
- [llvm] [NVPTX] Generalize register-or-immediate operands for integer instructions (PR #229920)
Durgadoss R via llvm-commits
- [llvm] e2a6fc5 - [BasicAA] Fix miscompilation with setjmp/longjmp due to missing longjmp re-entry paths in alias analysis (#212297)
via llvm-commits
- [llvm] [SLP][Experimental] Skip long runtime-check `or` chains in front of vectorized loops (PR #229800)
Igor Kirillov via llvm-commits
- [llvm] [IR] Introduce intrinsics to model allocator provenance semantics (PR #228048)
Antonio Frighetto via llvm-commits
- [llvm] ff423b9 - [SLP]Fix masked gather GEPs cost, skip GEP index seeds (#228918)
via llvm-commits
- [llvm] Transforms: Use changeToCall in LowerInvoke (PR #230055)
Matt Arsenault via llvm-commits
- [llvm] 1301afd - [LLVM][CodeGen][SVE] Improve lowering for v1f32/f64 when NEON is not available. (#229724)
via llvm-commits
- [llvm] [AArch64] Remove unused CHECK lines. NFC (PR #230058)
David Green via llvm-commits
- [llvm] a9a144e - [AArch64] Remove unused CHECK lines. NFC (#230058)
via llvm-commits
- [llvm] [AMDGPU][NFC] Explicitly narrow conversions in the assembly parser (PR #215202)
via llvm-commits
- [llvm] [VPlan] Drop add-induction nsw on smin-step (PR #230060)
Ramkumar Ramachandra via llvm-commits
- [flang] [llvm] [OpenMP] Introduce descriptors for directives, clause groups and sets (PR #229522)
Tom Eccles via llvm-commits
- [llvm] [VPlan] Add infrastructure for check-first early-exit vectorization. (PR #220162)
Arjun H Kumar via llvm-commits
- [llvm] ddd4d73 - [OpenMP] Fix debug locations in GPU reductions (#228622)
via llvm-commits
- [llvm] 2812ee7 - [AMDGPU] Extend new hazard CFG walk for wmma instruction support (#229428)
via llvm-commits
- [clang] [lld] [llvm] Machine outline fix for LTO + linker scripts (PR #230063)
Garvit Gupta via llvm-commits
- [llvm] [mlir][bazel] Build MLIR bindings extensions with `nanobind_bazel` (PR #169306)
Nicholas Junge via llvm-commits
- [llvm] PPC: Fix extsw elimination when the input reads a subregister (PR #208052)
Matt Arsenault via llvm-commits
- [clang] [lld] [llvm] [ConstantMerge] Merge constants within the same linker output section (PR #230065)
Garvit Gupta via llvm-commits
- [llvm] [GlobalISel] Add nomerge attribute support (PR #224332)
Milica Kovacevic via llvm-commits
- [llvm] [CodeGen] Ignore debug instructions in the VLIW block-size heuristic (PR #224497)
Yashas Andaluri via llvm-commits
- [llvm] [EarlyIfConversion] Add cascading conditional branches as a conversion pattern (PR #218415)
Jonathan Cohen via llvm-commits
- [llvm] [NFC][LV] Tests for parallel def-use chains (PR #225405)
Nashe Mncube via llvm-commits
- [llvm] [GlobalISel][TableGen] Add GIReplaceRegWithConstant/FConstant builtins (PR #222298)
Vikash Gupta via llvm-commits
- [llvm] [InstCombine] Shrink fast logarithm intrinsics to float (PR #228805)
Dávid Bolvanský via llvm-commits
- [llvm] [AArch64] Remove stale packed extend/shift for memory ops(NFC) (PR #230068)
Tomer Shafir via llvm-commits
- [llvm] 2296021 - [NFC][AMDGPU] Add tests for the cost of vector f32 to f16 fptrunc (#229949)
via llvm-commits
- [llvm] 55ea1f4 - [AMDGPU] Price vector f32 to f16 fptrunc by its packing form (#229950)
via llvm-commits
- [llvm] [WebAssembly] Preserve tail-call pseudos during stack-protector splitting (PR #230073)
Ben Silver via llvm-commits
- [llvm] [AMDGPU] Skip meta instructions in required export priority checks (PR #230074)
Yongqiang Tian via llvm-commits
- [clang] [llvm] [gcov] Add support for -fprofile-abs-path (PR #230077)
Nicolas Qiu Guichard via llvm-commits
- [llvm] [Offload] Remove unused entries from Plugin interface (PR #230078)
Alex Duran via llvm-commits
- [llvm] [X86] Don't fold fp_round pairs into f64->f16 with F16C (PR #230079)
Aochang Liu via llvm-commits
- [llvm] [AArch64] add patterns for SQABS (PR #229812)
Paul Walker via llvm-commits
- [llvm] 8e5a934 - CodeGen: Run LiveIntervals before PHIElimination and drop LiveVariables from it (#228618)
via llvm-commits
- [llvm] PHIElimination: Remove LiveVariables/LiveIntervals abstraction helpers (PR #228849)
Matt Arsenault via llvm-commits
- [llvm] [APFloat] Turn negative results in unsigned formats into NaN (PR #230086)
Hung-Kuan Tseng via llvm-commits
- [clang] [llvm] [APFloat] Signal overflow when rounding delivers the largest finite value (PR #220020)
Hung-Kuan Tseng via llvm-commits
- [llvm] [DAG] Consistently use Function argument in getRecipEstimate callbacks (PR #230088)
Simon Pilgrim via llvm-commits
- [llvm] RuntimeLibcalls: Pass the default calling convention to libraries (PR #230089)
Matt Arsenault via llvm-commits
- [llvm] [AMDGPU][NFC] Explicitly narrow conversions in the disassembler (PR #215193)
via llvm-commits
- [llvm] [LV] Remove manual "has getOpcode()" checks in Recipe_match (NFC) (PR #230095)
Benjamin Maxwell via llvm-commits
- [llvm] [CostModel][X86] Update AVX1/AVX2 uitofp v2i64 -> v2f32 costs based off worst case llvm-mca numbers (PR #230098)
Tim Besard via llvm-commits
- [llvm] [WIP][X86] Don't SimplifyDemandedVectorElts for and(X,0)/or(X,-1)/andnot(X,0) style patterns (PR #230099)
Simon Pilgrim via llvm-commits
- [llvm] b4badb6 - [AArch64] Add support for C2 CPUs (#229055)
via llvm-commits
- [llvm] [VPlan] Strip unnecessary WidenInd constructor (NFC) (PR #230100)
Ramkumar Ramachandra via llvm-commits
- [llvm] [LV] Tidy up early bail outs related to runtime checks (PR #228077)
Florian Hahn via llvm-commits
- [llvm] RegisterPressure: Detect dead physreg defs from LiveIntervals (PR #225079)
Matt Arsenault via llvm-commits
- [llvm] 0a920bb - [LV] Remove manual "has getOpcode()" checks in Recipe_match (NFC) (#230095)
via llvm-commits
- [llvm] [CodeGenPrepare] Handle negative GEP offsets in splitLargeGEPOffsets (PR #227627)
via llvm-commits
- [llvm] [DAGCombine] Combine non-extending partial.reduce.fadd into fma (PR #230120)
Graham Hunter via llvm-commits
- [llvm] [SLP][AArch64] Add tests for splats folded into indexed FMLA (NFC) (PR #230121)
Tim Besard via llvm-commits
- [llvm] [AMDGPU][NFC] Explicitly narrow conversions in subtarget and analysis passes (PR #215194)
via llvm-commits
- [llvm] [SLP]Emit combined nodes of reused gather PHI operands (PR #230122)
Alexey Bataev via llvm-commits
- [llvm] [NVPTX] Custom lower copysign operations (PR #230107)
via llvm-commits
- [llvm] fffc219 - [X86] Add smulh/umulh vector codegen tests (#230103)
via llvm-commits
- [llvm] [X86] Add smulh/umulh vector codegen tests (PR #230103)
Simon Pilgrim via llvm-commits
- [llvm] [AArch64] Preserve exceptions from discarded strict SVE FP operations (PR #230112)
via llvm-commits
- [llvm] [Offload][omp] Add iterate_platforms helper (PR #230144)
Alex Duran via llvm-commits
- [llvm] [AArch64] Restore always-inline behaviour for incompatible SME attributes. (PR #218727)
Benjamin Maxwell via llvm-commits
- [llvm] 64a5365 - [X86] Prefer v8i64 MULHS expansion on non-DQI AVX512 targets (#229704)
via llvm-commits
- [llvm] [AArch64] Add ext mask shuffle costs (PR #226136)
David Green via llvm-commits
- [llvm] d67ecfb - [RISCV] Fix violations of SDTCisSameAs when creating vector ISD nodes. (#228215)
via llvm-commits
- [llvm] [SLP][AArch64] Don't count splats folded into indexed FMLA in inst count (PR #230105)
Tim Besard via llvm-commits
- [llvm] Revert "workflows/precommit: Use single shared job to compute project… (PR #230124)
Aiden Grossman via llvm-commits
- [llvm] 9e01927 - [MLGO] Avoid NaN regalloc features for unspillable live ranges (#229751)
via llvm-commits
- [llvm] [ScheduleDAGFast] Fix crash due to incorrect live reg delay in fast scheduler (PR #200567)
Ye Tian via llvm-commits
- [llvm] a9d57c0 - [GISel][AArch64][Mips][RISCV] Allow G_BRINDIRECT to be selected via tablegen. (#229450)
via llvm-commits
- [llvm] [InstCombine] Fold disjoint or of add/xor with constants (PR #227497)
Alex MacLean via llvm-commits
- [llvm] [InstSimplify] Forbid undef folds in foldOrCommuteConstant when undef cannot be used. (PR #170670)
via llvm-commits
- [llvm] [GlobalISel] [AArch64] Support ROTR shift amount masking in SelectShiftMask (PR #226579)
Deepak Shirke via llvm-commits
- [llvm] fb71bb8 - [AArch64] Fold DUPLANE of 0 to zeroimm (#228415)
via llvm-commits
- [llvm] 3e6acc8 - [CodeGen] Add a FunctionSplittingMode target option (#226575)
via llvm-commits
- [llvm] c898380 - [RISCV][P-ext] Merge RISCVISD::PNCLIP(U)P and PNCLIP(U)P_W. (#228191)
via llvm-commits
- [llvm] [RISCV][P-ext] Merge RISCVISD::PNCLIP(U)P and PNCLIP(U)P_W. (PR #228191)
Craig Topper via llvm-commits
- [llvm] e839e46 - [RISCV] Change custom promotion of riscv_clmulh/r (#228846)
via llvm-commits
- [llvm] [AArch64] Use VectorInstrContext in getScalarizationOverhead. (PR #177201)
Florian Hahn via llvm-commits
- [llvm] f9c83e6 - [SandboxIR] Move IRBuilder from Context to Module (NFC) (#230033)
via llvm-commits
- [llvm] [LoopInterchange] Extract statically bounded outer epilogues (PR #224196)
Madhur Amilkanthwar via llvm-commits
- [llvm] bd3ce12 - Transforms: Use changeToCall in LowerInvoke (#230055)
via llvm-commits
- [llvm] [Offload][omp] Move some OpenMP env vars out of the plugins (PR #230140)
Robert Imschweiler via llvm-commits
- [llvm] [AArch64][CostModel] Consider some nxv1 casts as illegal (PR #230153)
Gaëtan Bossu via llvm-commits
- [llvm] [AArch64][CostModel] Allow more nxv1 operations on integers (PR #226146)
Gaëtan Bossu via llvm-commits
- [llvm] [Verifier][NFC] Merge negative intrinsic signature tests into a single file (PR #230160)
Yingwei Zheng via llvm-commits
- [llvm] 996ac2e - [CostModel][X86] Update AVX1/AVX2 uitofp v2i64 -> v2f32 costs based off worst case llvm-mca numbers (#230098)
via llvm-commits
- [llvm] [RISCV][RuntimeDyld] Add RV64 call stubs to RuntimeDyld (PR #229469)
Fu-Chiang Chang via llvm-commits
- [llvm] [DAG] Expand mask_beforefirst during promotion (PR #230163)
Luke Lau via llvm-commits
- [llvm] [LAA] Enable diff checks for non-unit constant stride (PR #188826)
Florian Hahn via llvm-commits
- [llvm] [X86] LowerSINT_TO_FP - pull out repeated CONCAT_VECTORS widening. NFC. (PR #230164)
Simon Pilgrim via llvm-commits
- [llvm] [XRay] Fix wrong entry sled placement after TRE (PR #230146)
Peter Arzt via llvm-commits
- [llvm] [Offload] Fix: Target JIT registers no backends when LLVM_LINK_LLVM_DYLIB=ON (PR #227370)
Joseph Huber via llvm-commits
- [llvm] [AMDGPU] Only legalize PHIs in SIFixSGPRCopies that need it (PR #226213)
Pablo Reble via llvm-commits
- [llvm] [VPlan] Keep VPPhi in ind, rec, rdx recipes (NFC) (PR #230168)
Ramkumar Ramachandra via llvm-commits
- [libcxx] [llvm] [libc++] Add CI running on macOS without exceptions (PR #230157)
Aiden Grossman via llvm-commits
- [llvm] [gn] port e65c4e339af01 (libc++ <debugging>) (PR #230170)
Nico Weber via llvm-commits
- [llvm] [RegisterCoalescer] Keep endpoints of lane-carrying IMPLICIT_DEF (PR #230141)
Vlad Korovin via llvm-commits
- [llvm] [gn build] Port 89c826d46383 (PR #230171)
Nico Weber via llvm-commits
- [llvm] [gn build] Port 89c826d463837 (PR #230172)
Nico Weber via llvm-commits
- [llvm] [gn build] Port 92f9852fc99b0 (PR #230173)
Nico Weber via llvm-commits
- [llvm] [gn build] Port b54c096369ab (PR #230174)
Nico Weber via llvm-commits
- [llvm] [DAGCombiner] Merge debug locations when combining stores (PR #225358)
via llvm-commits
- [llvm] [X86] Remove v8i64 S/UMUL_LOHI handling (PR #230206)
Simon Pilgrim via llvm-commits
- [llvm] [CodeGen] Fix assertion when discovering partial live-outs (PR #224323)
via llvm-commits
- [llvm] 6d56e63 - [gn] port 9f3af63e886d (#230199)
via llvm-commits
- [llvm] [gn] port 9f3af63e886d (PR #230199)
Nico Weber via llvm-commits
- [llvm] [DAGCombiner] Avoid misleading inline frames in store merging (PR #225358)
via llvm-commits
- [llvm] 3fab81a - [X86] LowerSINT_TO_FP - pull out repeated CONCAT_VECTORS widening. NFC. (#230164)
via llvm-commits
- [llvm] [Hexagon] Translate 6 more IEEE HVX intrinsics to QFloat (PR #227779)
Sumanth Gundapaneni via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) tranformations in BinOpSameOpcodeHelper (PR #181731)
Alexey Bataev via llvm-commits
- [llvm] [libsycl] Add handler counterparts to queue USM API (PR #230101)
Kseniya Tikhomirova via llvm-commits
- [llvm] [SROA] Improve select folding around casts (PR #222466)
Daniel Donenfeld via llvm-commits
- [llvm] [AArch64] Add NVCAST to inputs of AArch64ISD::REV32/REV64 to match SDTypeProfile. (PR #230211)
Craig Topper via llvm-commits
- [llvm] [libc++] Update macOS CI jobs to C++26 (PR #230136)
Nikolas Klauser via llvm-commits
- [clang] [llvm] [openmp] [CLANG][OpenMP] Restore loop variable values after transformation directives (PR #214717)
Zahira Ammarguellat via llvm-commits
- [llvm] CodeGen: Drop dead LiveVariables optional-analysis paths (PR #230147)
Matt Arsenault via llvm-commits
- [llvm] [SPIRV] Extend function decoration test coverage (PR #230180)
Marcos Maronas via llvm-commits
- [clang] [flang] [llvm] [Passes] Port -print-pipeline-passes to PassesOptions (PR #230220)
Alexis Engelke via llvm-commits
- [llvm] Revert "[AMDGPU] Reserve ENABLE_WAVEFRONT_SIZE32 on gfx125x" (PR #230228)
Chinmay Deshpande via llvm-commits
- [llvm] [X86][AVX-512] Fold `zext(and(bitcast(mask), C))` --> `and(anyext(bitcast(mask)), zext(C))` (PR #220251)
via llvm-commits
- [llvm] bcad250 - [SPIRV] Extend function decoration test coverage (#230180)
via llvm-commits
- [clang] [llvm] Enable ClangIR Build By Default (PR #221291)
Erich Keane via llvm-commits
- [llvm] [SPIRV] Support UserTypeGOOGLE decorations (PR #230242)
Alexander Shaposhnikov via llvm-commits
- [llvm] [NVPTX] Select mul.wide.u64 based on known bits (PR #228285)
via llvm-commits
- [llvm] 3bcdf20 - [llubi] Avoid overflow when computing EVL with large VF (#230231)
via llvm-commits
- [llvm] [llubi] Avoid overflow when computing EVL with large VF (PR #230231)
Yingwei Zheng via llvm-commits
- [llvm] [GlobalISel] Port simplifyDemandedBits to CombinerHelper (PR #198808)
Alan Li via llvm-commits
- [llvm] [BOLT][AArch64] Don't treat entry mapping symbols as function symbols (PR #230244)
Amir Ayupov via llvm-commits
- [lldb] [llvm] [lldb] Split register descriptions from ProcessUtility (PR #222330)
Alex Langford via llvm-commits
- [llvm] [AMDGPU] Use instrLatency for memory HWUI cycle accounting (PR #205418)
Jeffrey Byrnes via llvm-commits
- [llvm] [CodeGen] Honor -function-splitting=none in BasicBlockSections (PR #227880)
Rahman Lavaee via llvm-commits
- [llvm] [RISCV] Inline RISCVAsmPrinter::emitAttributes. NFC (PR #230235)
Sam Elliott via llvm-commits
- [flang] [llvm] [flang] Round COMPLEX(4)**INTEGER to single precision once (PR #230257)
Chris Frisz via llvm-commits
- [clang] [flang] [llvm] [clang][flang][OpenMP] Fix context selector matching and scoring (PR #224431)
via llvm-commits
- [llvm] 20738ea - CodeGen: Drop dead LiveVariables optional-analysis paths (#230147)
via llvm-commits
- [lld] [llvm] CodeGen: Drop dead LiveVariables optional-analysis paths (PR #230147)
Matt Arsenault via llvm-commits
- [lld] [llvm] CodeGen: Strip LiveVariables down to dead flag computation (PR #230149)
Matt Arsenault via llvm-commits
- [llvm] 7208ba2 - [RISCV] Only expand lla to qc.e.li in .option nopic (#229995)
via llvm-commits
- [llvm] CodeGen: Strip LiveVariables down to dead flag computation (PR #230149)
Matt Arsenault via llvm-commits
- [llvm] [llvm][IRMover/TypeFinder] Also visit debug info from functions/global variables/instructions (PR #230236)
Arthur Eubanks via llvm-commits
- [llvm] [DAGCombiner] - draft - remove redundant subs in X - (trunc X - B) if X's value fits it the low half of X (PR #230277)
Yoonseo Choi via llvm-commits
- [llvm] [NVPTX] Fold vector and 16 bit float FMAs in the IR peephole pass (PR #224018)
Daniel Donenfeld via llvm-commits
- [llvm] [Bazel] Fixes a789eb1 (PR #230272)
Aiden Grossman via llvm-commits
- [clang] [llvm] [Clang][X86] Resolve target_clones calls through GlobalOpt (PR #230278)
Oleksandr Kulkov via llvm-commits
- [llvm] [SCEV] Make operand complexity order total. (PR #230252)
Florian Hahn via llvm-commits
- [llvm] 660a0e6 - [NFC][SLP] Precommit tests for the phantom load saving (#228901)
via llvm-commits
- [llvm] [AMDGPU] Limit the fmul fusion discount to a matching context type (PR #229845)
Dmitry Sidorov via llvm-commits
- [llvm] [TTI][SLP] Tell the target when the users of a cost context stay scalar (PR #228902)
Dmitry Sidorov via llvm-commits
- [llvm] [Offload] Add OFFLOAD_FORCE_SYNC_OPS force-synchronization escape hatch (PR #222635)
Jan Patrick Lehr via llvm-commits
- [llvm] ae2fee0 - AMDGPU/GlobalISel: Legalize bf16 g_fcmp (#227909)
via llvm-commits
- [llvm] [LAA] Fix missing RT checks for unknown deps via non-header phis. (PR #230292)
Florian Hahn via llvm-commits
- [llvm] X86: Only drop EFLAGS def in convertToThreeAddress if MI defines EFLAGS (PR #230293)
Matt Arsenault via llvm-commits
- [llvm] [LFI][X86] Add memory access rewrites (PR #227511)
Nick Desaulniers via llvm-commits
- [llvm] [AMDGPU] Move to new pass spelling virt-reg-rewriter in 2 tests (PR #230287)
Matt Arsenault via llvm-commits
- [llvm] 7e1fa69 - [AMDGPU] Move to new pass spelling virt-reg-rewriter in 2 tests (#230287)
via llvm-commits
- [llvm] f9c8a1f - CodeGen: Strip LiveVariables down to dead flag computation (#230149)
via llvm-commits
- [llvm] LiveVariables: Remove dead live-in handling and Defs plumbing (PR #230150)
Matt Arsenault via llvm-commits
- [llvm] [SLP][NFC]Add a test with missing vectorization, NFC (PR #230295)
Alexey Bataev via llvm-commits
- [llvm] b9dd495 - [SLP][NFC]Add a test with missing vectorization, NFC
via llvm-commits
- [llvm] [InstCombine] Mark profiles unknown in foldClampRangeOfTwo (PR #230297)
Aiden Grossman via llvm-commits
- [llvm] 974e95f - [CodeGen] Force compiler to generate default destructor for class (#229944)
via llvm-commits
- [llvm] [RISC-V][MC] Add test for C/Zce implication in .option arch (PR #230298)
Alexander Richardson via llvm-commits
- [llvm] [llvm-profgen] Add ITM data trace support (PR #230309)
via llvm-commits
- [llvm] [llvm-profgen] Add ITM data trace support (PR #230310)
via llvm-commits
- [llvm] [orc-rt] Mark LockedAccess's constructor as noexcept. (PR #230311)
Lang Hames via llvm-commits
- [llvm] [InstCombine] Share common binop folds between icmp and [su]cmp (PR #222978)
Kenzo Tjandra via llvm-commits
- [llvm] f8f362d - [AMDGPU] Use vector memory types for 64/128-bit cooperative atomics (#229501)
via llvm-commits
- [llvm] [AMDGPU] Promote wmma_f32_16x16x128_f8f6f4 to scaled version on gfx1250-strict (PR #230253)
Siu Chi Chan via llvm-commits
- [llvm] [BOLT] Keep ambiguous references next to function boundaries valid (PR #230321)
Amir Ayupov via llvm-commits
- [llvm] [JITLink][PowerPC] Add ELF PPC32 support (PR #230262)
Matt Turner via llvm-commits
- [llvm] [gn build] Port a789eb1d665b (PR #230326)
Nico Weber via llvm-commits
- [llvm] 8b7a4fb - [gn build] Port a789eb1d665b (#230326)
via llvm-commits
- [llvm] [LoongArch] Ignore debug uses when folding large offsets (PR #230290)
via llvm-commits
- [compiler-rt] [sanitizer_common][Darwin] Add debug memory region support for iOS 27.0 (PR #217527)
Andrew Haberlandt via llvm-commits
- [llvm] [AMDGPU] NFC: Drop constexpr from getFlavorName and getCoExecMask (PR #230334)
Austin Kerbow via llvm-commits
- [llvm] [GlobalISel] Take into account the address space when calculating memcpy offsets (PR #230216)
Daniel Paoliello via llvm-commits
- [llvm] [VectorCombine] Fold select chain to and/or using freezes (PR #230337)
via llvm-commits
- [llvm] [RISCV] Avoid early split for fixed-length VECTOR_DEINTERLEAVE (PR #230339)
via llvm-commits
- [llvm] [Support] Disable CFI on typed Allocate to unconstructed storage (PR #217992)
via llvm-commits
- [llvm] [RISCV] Reject reserved cm.mvsa01/cm.mva01s encodings in the disassembler (PR #229091)
Zeyi Xu via llvm-commits
- [llvm] [NVPTX] Cleanup support for atomic loads/stores (PR #201468)
Akshay Deodhar via llvm-commits
- [llvm] [RISCV] Reject reserved (qc.)cm.mvsa01/(qc.)cm.mva01s encodings in the disassembler (PR #229091)
Zeyi Xu via llvm-commits
- [llvm] [RISCV] Correct the CFA offsets for the stack probe loop. (PR #230322)
Pengcheng Wang via llvm-commits
- [llvm] [Mips] Fix R6 floating-point condition register handling (PR #226838)
Ying Huang via llvm-commits
- [llvm] [RISCV] Ignore debug uses in vector peephole folds (PR #230319)
Pengcheng Wang via llvm-commits
- [llvm] [SPIRV] Infer external pointer parameter types from mangled names (PR #230342)
Alexander Shaposhnikov via llvm-commits
- [llvm] [KnownFPClass] Add support for non-intrinsic math libcalls (PR #216234)
Eric Ross via llvm-commits
- [clang] [llvm] [LLVMABI][AARCH64] Implement Pure Scalable Type handling (PR #227504)
Madhur Amilkanthwar via llvm-commits
- [llvm] [AMDGPU] Avoid wmma bank conflict for GFX11 and GFX12 (PR #205530)
via llvm-commits
- [llvm] be2daae - [RISCV] Reject reserved (qc.)cm.mvsa01/(qc.)cm.mva01s encodings in the disassembler (#229091)
via llvm-commits
- [llvm] [Hexagon] Require asserts for autohvx xqf-postra-conv-double2.ll (PR #230348)
Brian Cain via llvm-commits
- [llvm] [AMDGPU] Insert MFMA anti-hints in GCNPreRAOptimizations (PR #218075)
Syadus Sefat via llvm-commits
- [llvm] [VectorCombine] Count the shuffle as removed only when its binop is removed in foldShuffleOfBinops (PR #228473)
via llvm-commits
- [llvm] [VectorCombine] Merge permutes of the same source in foldShuffleOfBinops (PR #228472)
via llvm-commits
- [llvm] [MemProf] Duplicate pruned (trimmed) contexts onto callee clones (PR #230351)
David Li via llvm-commits
- [llvm] [AMDGPU] add v_cndmask_b16 to macro funsion (PR #230354)
Guo Chen via llvm-commits
- [llvm] [RISCV] Remove RISCVISD::PPAIROE_H. NFC (PR #230356)
Craig Topper via llvm-commits
- [llvm] 7dd738d - [NFC][RegisterPressure] Add RegisterOperands::restoreLivenessFlags helper (#229952)
via llvm-commits
- [llvm] [NFC][RegisterPressure] Add RegisterOperands::restoreLivenessFlags helper (PR #229952)
Puyan Lotfi via llvm-commits
- [llvm] [RISCV][P-ext] Combine PPAIR isel patterns by using HwMode types. (PR #230357)
Craig Topper via llvm-commits
- [llvm] c0a1a2c - [AMDGPU] Promote wmma_f32_16x16x128_f8f6f4 to scaled version on gfx1250-strict (#230253)
via llvm-commits
- [llvm] [ProfCheck] Enable PhaseOrdering and mark invented selects unknown (PR #230360)
Alok Kumar Sharma via llvm-commits
- [llvm] [NFC][llvm-ar] Restore file permissions in error-opening-permission.test (PR #230361)
Vitaly Buka via llvm-commits
- [llvm] [SLP][modularisation][NFC] Move isFirstInsertElement/getDebugLocFromPHI to SLPUtils (PR #222612)
Madhur Amilkanthwar via llvm-commits
- [llvm] 15e2ad9 - [cmake] Limit `/Zc:dllexportInlines-` flag to C/C++ (#227791)
via llvm-commits
- [llvm] [ids-check] Print diff before updating issue (PR #230315)
Fabrice de Gans via llvm-commits
- [llvm] [RISCV][P-ext] Combine duplicate mhacc(su)_h_b(0/1) patterns. NFC (PR #230358)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Combine MHACC_H_B0 and MHACC_W_H0. NFC (PR #230362)
Craig Topper via llvm-commits
- [llvm] 8c4c52b - [AMDGPU] Insert MFMA anti-hints in GCNPreRAOptimizations (#218075)
via llvm-commits
- [llvm] [AMDGPU] Insert wmma coexec aware anti-hints rules (PR #226397)
Syadus Sefat via llvm-commits
- [llvm] [InstCombine] Fold select of low-bit icmp to and (PR #230363)
via llvm-commits
- [llvm] [Hexagon] Test requires asserts (PR #230364)
Alina Sbirlea via llvm-commits
- [llvm] 16dae0f - [Passes] Port -print-pipeline-passes to PassesOptions (#230220)
via llvm-commits
- [flang] [llvm] [CodeGen] Merge MachineFunctionSplitter into BasicBlockSections (PR #230365)
Rahman Lavaee via llvm-commits
- [llvm] b1c00b4 - [Hexagon] Require asserts for autohvx xqf-postra-conv-double2.ll (#230348)
via llvm-commits
- [llvm] [RISCV][P-ext] Combine RISCVISD::MQACC_00 and PMQACC_W_H00. NFC (PR #230368)
Craig Topper via llvm-commits
- [llvm] [LoongArch] Add scheduler definitions for LoongArch with schedule model for LA464/LA664 (PR #226064)
via llvm-commits
- [llvm] 8ab8809 - [XRay] Fix wrong entry sled placement after TRE (#230146)
via llvm-commits
- [llvm] [Scalar] Declare command line options in TableGen (PR #230370)
Fangrui Song via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (PR #230377)
Craig Topper via llvm-commits
- [llvm] [LSR] Count unfolded adds of ICmpZero formulae and negations in outer loops (PR #230118)
Tim Besard via llvm-commits
- [llvm] [LSR] Break ties between offset winner registers by instruction count (PR #230119)
Tim Besard via llvm-commits
- [llvm] [AMDGPU][NFC] Fix bogus register names in AGPR rewrite debug output (PR #230376)
Matt Arsenault via llvm-commits
- [llvm] [AArch64] Fix SDAG lowering of small ints on the stack (PR #230142)
David Green via llvm-commits
- [llvm] [InstCombine] Don't narrow FP ops when the destination range is too small (PR #230379)
via llvm-commits
- [llvm] LiveVariables: Only visit tracked physical registers (PR #230151)
Matt Arsenault via llvm-commits
- [llvm] be4711c - AMDGPU: Use functions in more operation tests instead of kernel loads (#230195)
via llvm-commits
- [llvm] AMDGPU: Use functions in more operation tests instead of kernel loads (PR #230195)
Matt Arsenault via llvm-commits
- [llvm] 47c2af9 - X86: Only drop EFLAGS def in convertToThreeAddress if MI defines EFLAGS (#230293)
via llvm-commits
- [llvm] RuntimeLibcalls: Organize functions into libraries (PR #229562)
Matt Arsenault via llvm-commits
- [llvm] [SampleProf] Accumulate vtable type counts for FS-discriminator-collapsed locations (PR #230380)
Mingming Liu via llvm-commits
- [llvm] CodeGen: Avoid re-sorting idx2MBBMap when inserting a block into SlotIndexes (PR #230374)
Nikita Popov via llvm-commits
- [llvm] [AMDGPU] Remove unused functions (NFC) (PR #230382)
Kazu Hirata via llvm-commits
- [llvm] AMDGPU: Fix true16 build_vector (0, x) pattern using a 16-bit shift operand (PR #230383)
Matt Arsenault via llvm-commits
- [llvm] [CodeGen] Define ~LazyMachineBlockFrequencyInfoPass out of line (PR #230384)
Fabrice de Gans via llvm-commits
- [llvm] [Support] Use a 64-bit byte count in SHA1 and SHA256 (PR #230385)
Buyun Xu via llvm-commits
- [libclc] [llvm] [flang-rt][libclc] Install multilib variants in their own directory (PR #230251)
Wenju He via llvm-commits
- [llvm] [SampleProf] Don't re-read profiles that are already loaded in incremental reads (PR #230386)
Mingming Liu via llvm-commits
- [llvm] [Offload][omp] Get Plugin active count and Device UID through liboffload (PR #230109)
Alex Duran via llvm-commits
- [llvm] [SPIR-V] Lower llvm.minimum/maximum instead of selecting fmin/fmax (PR #225667)
Tim Besard via llvm-commits
- [libc] [llvm] [libc][CI] Add CMake format check to code-format-helper (PR #228526)
Jeff Bailey via llvm-commits
- [llvm] 4ef09f6 - CodeGen: Avoid re-sorting idx2MBBMap when inserting a block into SlotIndexes (#230374)
via llvm-commits
- [llvm] CodeGen: Check liveness before PHI sources when splitting a critical edge (PR #230375)
Matt Arsenault via llvm-commits
- [llvm] f72b2a3 - [libc][CI] Add CMake format check to code-format-helper (#228526)
via llvm-commits
- [llvm] [offload][nfc] Pull OpenMP's InteropTbl out of PluginManager (PR #230283)
Alex Duran via llvm-commits
- [llvm] [InstCombine] Reuse existing sub result when folding "A - (B + C)" (PR #230387)
via llvm-commits
- [clang] [llvm] [SystemZ, Clang] Enable verification of int args extensions by default. (PR #223755)
Nikita Popov via llvm-commits
- [clang] [clang-tools-extra] [lld] [lldb] [llvm] [mlir] [SystemZ][z/OS] Mark text files as text (PR #228437)
James Henderson via llvm-commits
- [llvm] X86: Fix missing requires asserts in test added by #230293 (PR #230389)
Matt Arsenault via llvm-commits
- [llvm] 33b4f3c - X86: Fix missing requires asserts in test added by #230293 (#230389)
via llvm-commits
- [llvm] d29c2e8 - [LV] Tidy up early bail outs related to runtime checks (#228077)
via llvm-commits
- [llvm] [AMDGPU] No s_delay_alu for VALU fast-forwarded operands (PR #230390)
Reem Elkhouly via llvm-commits
- [llvm] 3eb0f76 - LiveVariables: Only visit tracked physical registers (#230151)
via llvm-commits
- [llvm] 7b0397a - [AMDGPU] Remove unused functions (NFC) (#230382)
via llvm-commits
- [clang] [llvm] [Clang][WebAssembly] Add wasm_global declaration attribute (PR #221909)
Changqing Jing via llvm-commits
- [llvm] 488f609 - X86: Mark EFLAGS clobbers dead in global base register setup (#229337)
via llvm-commits
- [clang] [llvm] [PowerPC] Prefix option names with ppc-, turn disable- to positive options (PR #230393)
Fangrui Song via llvm-commits
- [llvm] [InstCombine] Don't propagate ninf when hoisting fneg above fmul/fdiv (PR #228671)
Divyansh Yadav via llvm-commits
- [llvm] 7a03c7f - [X86] Ignore debug instructions in FixupLEAs search distance (#222871)
via llvm-commits
- [llvm] [X86] Ignore debug instructions in FixupLEAs search distance (PR #222871)
Jeremy Morse via llvm-commits
- [compiler-rt] [asan] Fix `unknown-crash` being reported for multi-byte errors, and incorrect memory access addresses being reported (PR #144480)
via llvm-commits
- [llvm] [Sandbox] Explicitly delete copy for Context (PR #230394)
Fabrice de Gans via llvm-commits
- [llvm] [SDAG] Look through multiple ADDs to compute offset in InferPointerInfo (PR #230395)
Pierre van Houtryve via llvm-commits
- [llvm] [RISCV][GlobalISel] Support orc.b (PR #230401)
Jian Li via llvm-commits
- [llvm] [ProfileData] Validate indexed profile header offsets and summary size (PR #230403)
Cyrus Ding via llvm-commits
- [llvm] [LAA] Avoid unprofitable unit-stride versioning (PR #225331)
Boyao Wang via llvm-commits
- [llvm] Revert "X86: Mark EFLAGS clobbers dead in global base register setup" (PR #230404)
Ingo Müller via llvm-commits
- [llvm] a150791 - Revert "X86: Mark EFLAGS clobbers dead in global base register setup" (#230404)
via llvm-commits
- [llvm] bf587de - [SimplifyCFG] Widen `!range` over PassThru in hoistConditionalLoadsStores (#217821)
via llvm-commits
- [llvm] 44a08dc - MC: Error if target did not register null streamer (#230024)
via llvm-commits
- [llvm] [VPlan] Allow a VPRecipeValue to carry a widened result type (NFC). (PR #228251)
Gaëtan Bossu via llvm-commits
- [llvm] [SimplifyCFG] Bypass PHI-only successors with equal incoming values (PR #230411)
Nopileos Isemados via llvm-commits
- [llvm] [ARM] Add combine for vector_deinterleave load (PR #225747)
Kamlesh Kumar via llvm-commits
- [llvm] [llvm] Remove incorrect IRMover assert (PR #221758)
Sebastian Neubauer via llvm-commits
- [llvm] [InstCombine] Fold and-of-srem-sign-contradictions to false (PR #230413)
via llvm-commits
- [llvm] [LV][NFC] Add early bail-out in setCostBasedWideningDecision (PR #230414)
David Sherwood via llvm-commits
- [llvm] [RegisterCoalescer] Only rewrite debug uses of the commuted value (PR #230416)
Aochang Liu via llvm-commits
- [llvm] [AMDGPU][GISel] Add regbank-aware same_val_zero fold to AMDGPURegBankCombiner (PR #225138)
via llvm-commits
- [llvm] Reapply "X86: Mark EFLAGS clobbers dead in global base register setup" (#230404) (PR #230418)
Matt Arsenault via llvm-commits
- [llvm] [Offload] Fix debug build after #222635 (PR #230421)
Alex Duran via llvm-commits
- [llvm] 51cc989 - [GVN] Do not create/destroy a local copy of GVNPassImpl in the run method (#229833)
via llvm-commits
- [llvm] [llvm-profgen] Generate static data-access profiles from AMD IBS and Intel PEBS samples in llvm-profgen (PR #230423)
chandan singh via llvm-commits
- [libc] [llvm] [libc] Files now flush on exit (PR #230215)
Pavel Labath via llvm-commits
- [llvm] c258a69 - [MLGO] Switch regalloc log context once per function (#229453)
via llvm-commits
- [llvm] [C API] Add LLVMIntrinsicGetOverloadTypes (PR #230425)
Valentin Churavy via llvm-commits
- [llvm] [InstCombine] Fold special cases of vector partial reduction add (PR #226714)
via llvm-commits
- [llvm] Test: Make the load/store alloca to volatile to stop promotion (PR #230426)
via llvm-commits
- [llvm] 4bdfa6e - [clang][NVPTX] Add overloaded fmul intrinsics (#224546)
via llvm-commits
- [llvm] [FuncSpec] Relax restrictions on code size growth (PR #196570)
Momchil Velikov via llvm-commits
- [llvm] [PassBuilder] Remove LoopFullUnrollPass from before LoopVectorize in the FullLTO pipeline (PR #197176)
Momchil Velikov via llvm-commits
- [clang] [llvm] Disable pre-link LTO vectorisation and loop unroll for at O2 and O3 optimisation levels (PR #195886)
Momchil Velikov via llvm-commits
- [llvm] [X86] concatSubVectors helper - use POISON as INSERT_SUBVECTOR base instead of UNDEF (PR #230427)
Simon Pilgrim via llvm-commits
- [llvm] [LoopDistribute] Statement-seeded partitioning for inner and outer loops (PR #230428)
Madhur Amilkanthwar via llvm-commits
- [clang] [llvm] [SSP] Add llvm.ssp.protected intrinsic for QualType-based stack protection (PR #183623)
Kuba Mracek via llvm-commits
- [llvm] Make tail calls work properly for Arm64 Windows (PR #230429)
Alastair Houghton via llvm-commits
- [llvm] 3f1aba5 - Reapply "X86: Mark EFLAGS clobbers dead in global base register setup" (#230404) (#230418)
via llvm-commits
- [clang] [llvm] [clang][X86] Reject register variables bound to status registers (PR #227576)
Akash Manna via llvm-commits
- [llvm] [InstCombine] Fold zero-test of "(X + Y) || Y" to "X || Y" (PR #230431)
via llvm-commits
- [llvm] Fix Attributor crash for unknown call return (PR #230432)
Changqing Jing via llvm-commits
- [llvm] [RegisterCoalescer] Try every commute partner that reads the copy destination (PR #230433)
Aochang Liu via llvm-commits
- [llvm] 3756ef8 - [Sandbox] Explicitly delete copy for Context (#230394)
via llvm-commits
- [llvm] [LV] Rewrite Scalable preference computation (NFCI) (PR #230182)
Paul Walker via llvm-commits
- [llvm] [AArch64][GlobalISel] Combine load+shuffles to LD2/3/4 (PR #230438)
David Green via llvm-commits
- [llvm] [DebugInfo] Print DWARF expressions without a DWARFUnit dependency (PR #229796)
Charles Zablit via llvm-commits
- [llvm] 3820435 - [SLP][NFC]Add a test with incorrect flag propagation, NFC
via llvm-commits
- [llvm] [SLP][NFC]Add a test with incorrect flag propagation, NFC (PR #230448)
Alexey Bataev via llvm-commits
- [llvm] [LV][REVEC] Initial support for re-vectorisation (PR #208213)
Gaëtan Bossu via llvm-commits
- [flang] [llvm] [flang] Classify IEEE_ARITHMETIC elemental functions and operators as SIMPLE (PR #223254)
Eugene Epshteyn via llvm-commits
- [llvm] [AMDGPU] Fold inlinable constants with V_FMAMK into VOP3 V_FMA (PR #230452)
Carl Ritson via llvm-commits
- [llvm] [SLP]Fix wrap flags for lanes emitted in negated form (PR #230453)
Alexey Bataev via llvm-commits
- [llvm] 1c10209 - [SLP]Fix wrap flags for lanes emitted in negated form
via llvm-commits
- [llvm] [BOLT][AArch64] Use full layout for clustered relaxation (PR #225728)
Adam Bzowski via llvm-commits
- [llvm] [LAA] Use EarliestEscapeAnalysis for alias queries. (PR #228179)
Florian Hahn via llvm-commits
- [llvm] [llvm][UnicodeData] Check for LIBXML_READER_ENABLED before building UnicodeCharSetsGenerator (PR #230254)
Corentin Jabot via llvm-commits
- [llvm] [llvm-mca][ARM] Fix Cortex-A9 dispatch hazards and NEON memory resources (PR #230454)
via llvm-commits
- [llvm] [ARM][Disassembler] Add missing predicates when decoding Thumb barriers (PR #223540)
via llvm-commits
- [llvm] [DebugInfo][SelectionDAG] Give bool parameters a location at function entry (PR #230457)
via llvm-commits
- [llvm] [VPlan] Rework mergeBlocksIntoPredecessors (NFC) (PR #230460)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SLP][NFC]Add a test with incorrect optimization, NFC (PR #230462)
Alexey Bataev via llvm-commits
- [llvm] 538cec9 - [SLP][NFC]Add a test with incorrect optimization, NFC
via llvm-commits
- [llvm] 9368462 - [AArch64] Use VectorInstrContext in getScalarizationOverhead. (#177201)
via llvm-commits
- [llvm] [AMDGPU][NFC] Explicitly narrow conversions in asm printing and MIR formatting (PR #215195)
via llvm-commits
- [llvm] [AMDGPU] Widen scalar bf16 minimumnum and maximumnum to packed ops (PR #230467)
Tim Gymnich via llvm-commits
- [llvm] [MLGO] Do not evict live ranges with the same cascade number (PR #230468)
Arseniy Obolenskiy via llvm-commits
- [llvm] [DebugInfo] Filter discarded ELF ranges when selecting a CU (PR #230469)
Longjun Luo via llvm-commits
- [llvm] [LLVM][CodeGen][SVE] Add lowering for bfloat strict-fp arithmetic operations. (PR #230470)
Paul Walker via llvm-commits
- [llvm] [AArch64][GlobalISel] Ignore debug instructions when reusing carry. (PR #228128)
Mihai Zegheru via llvm-commits
- [llvm] 7a005aa - [ARM][Disassembler] Add missing predicates when decoding Thumb barriers (#223540)
via llvm-commits
- [llvm] fefeef2 - [IR] Name unnamed values after each pass (#221519)
via llvm-commits
- [llvm] Revert "[DAG] Expand mask_beforefirst during promotion" (PR #230471)
Luke Lau via llvm-commits
- [llvm] [AArch64][llvm] Remove redundant code in tryParseSVEPattern() (PR #230043)
via llvm-commits
- [llvm] [mlir] [MLIR][Remark] Report why the LLVM remark streamer could not be created (PR #230130)
Henrich Lauko via llvm-commits
- [llvm] [RISCV] Generalize combineBinOpOfZExt to allow sext (PR #230472)
Luke Lau via llvm-commits
- [llvm] [SCEVExp] Only count instructions needing analysis in reuse budget. (PR #230446)
Nikita Popov via llvm-commits
- [llvm] [SLP] Do not copy i1 reduction flags to booleanized reduction opsFix 230419 slp booleanized disjoint (PR #230474)
Aochang Liu via llvm-commits
- [llvm] [AArch64][GlobalISel] Remove G_DUP, replacing it with G_SPLAT_VECTOR (PR #230475)
David Green via llvm-commits
- [llvm] CodeGen: Only visit live-out vregs when splitting a critical edge with LIS (PR #230441)
Nikita Popov via llvm-commits
- [llvm] [LV] Improve code around inductions, reductions (NFC) (PR #230476)
Ramkumar Ramachandra via llvm-commits
- [llvm] [IndVarSimplify] Ensure fp to int32 conversion stays in bound (PR #230179)
Antonio Frighetto via llvm-commits
- [llvm] [NVPTX] Generalize flexible operands to floating-point instructions (PR #230258)
Durgadoss R via llvm-commits
- [llvm] [SDAG] Add MLOAD/MSTORE mask scalarization support to LegalizeVectorTypes (PR #230202)
Mattéo Rizza Murgier via llvm-commits
- [llvm] [ConstraintElim] Divide rows by the GCD of their coefficients. (PR #228765)
Florian Hahn via llvm-commits
- [llvm] 35eee97 - [SCEVExp] Only count instructions needing analysis in reuse budget. (#230446)
via llvm-commits
- [llvm] 43c3bf2 - CodeGen: Only visit live-out vregs when splitting a critical edge with LIS (#230441)
via llvm-commits
- [llvm] [AArch64] Reduce AddedComplexity value on atomic hint patterns (NFC) (PR #230478)
Kerry McLaughlin via llvm-commits
- [llvm] [Support] Fix SpecificBumpPtrAllocator move assignment leak (PR #230480)
via llvm-commits
- [clang] [llvm] [AArch64] Implement the atomic store with hint intrinsic (PR #198316)
Kerry McLaughlin via llvm-commits
- [llvm] 07dc962 - [X86] Declare memory effects for the AMX intrinsics (#229025)
via llvm-commits
- [llvm] [offload][nfc] Use the PluginManager instance instead of the global PM (PR #230481)
Ivan R. Ivanov via llvm-commits
- [llvm] [RISCV] Custom lower fixed length mask_beforefirst (PR #230482)
Luke Lau via llvm-commits
- [llvm] Fix overflow issues associated with shufflevector for landing PR #229662 (PR #230410)
Paul Walker via llvm-commits
- [llvm] [LV] Add helpers to extend VPWidenMemIntrinsicRecipe to non-VP intrinsics (NFC) (PR #230486)
Benjamin Maxwell via llvm-commits
- [llvm] 2acc171 - [AArch64][CostModel] Consider some nxv1 casts as illegal (#230153)
via llvm-commits
- [llvm] [PhaseOrdering][X86] vector-reduction-known-first-value.ll - add x86-64/x86-64-v2/x86-64-v3/x86-64-v4 test coverage (PR #230506)
Simon Pilgrim via llvm-commits
- [llvm] [X86] LowerMSTORE - fix copy+paste assert message typo. NFC. (PR #230507)
Simon Pilgrim via llvm-commits
- [llvm] [SCEV] Scale canReuseInstruction's walk budget with the expression size. (PR #230508)
Florian Hahn via llvm-commits
- [llvm] 2e61419 - [ARM][AArch64] Introduce the Armv9.8-A architecture version (#229361)
via llvm-commits
- [clang] [llvm] [AArch64][llvm] Improve diagnostics for invalid operands (PR #230436)
Jonathan Thackray via llvm-commits
- [llvm] a8f1f81 - [AArch64][llvm] Armv9.8-A: Add support for FEAT_CFLT (Conditional Fault instructions) (#229362)
via llvm-commits
- [llvm] [AArch64][llvm] Improve diagnostics for invalid operands (PR #230436)
Jonathan Thackray via llvm-commits
- [llvm] f24645c - [AArch64][llvm] Armv9.8-A: Add support for FEAT_LSC64B (single-copy atomic 64-byte load/store) (#229363)
via llvm-commits
- [llvm] 4dc2239 - [AArch64][llvm] Armv9.8-A: Add support for FEAT_RLCS (release consistency scoping) (#229364)
via llvm-commits
- [llvm] bd58c85 - [AArch64][llvm] Fix constant expressions in CMPBR immediate aliases (#230050)
via llvm-commits
- [llvm] [ValueTracking] Look through constant exprs in aggregate constants (PR #230487)
via llvm-commits
- [llvm] e0a8d72 - [PhaseOrdering][X86] vector-reduction-known-first-value.ll - add x86-64/x86-64-v2/x86-64-v3/x86-64-v4 test coverage (#230506)
via llvm-commits
- [llvm] [BOLT][AArch64] Shrink large binary call relaxation test (PR #230526)
Paul Osmialowski via llvm-commits
- [clang] [llvm] [IR] Instruction::maySynchronize should be false for monotonic atomics. (PR #230299)
Nikita Popov via llvm-commits
- [llvm] [llubi] Reject invalid types (PR #230495)
via llvm-commits
- [llvm] [X86] Add lowering for saturating casts of float vectors to integers (PR #230537)
Jörn Horstmann via llvm-commits
- [llvm] [IR] Fix InstNamer test failure under profile checking (PR #230527)
Yaxun Liu via llvm-commits
- [llvm] [SPIR-V] Type the get_fence pointer argument as i8 (PR #230541)
Arseniy Obolenskiy via llvm-commits
- [llvm] [lit] Allow to filter tests when using `--time-trace-output` (PR #228523)
Eric Miotto via llvm-commits
- [llvm] [InstCombine] Fix infinite loop in demanded FP-class simplification (PR #230544)
via llvm-commits
- [llvm] [LAA] Increase vectorize memory check threshold to 192 (PR #228382)
Igor Kirillov via llvm-commits
- [llvm] f5bd08d - [AArch64] Reduce AddedComplexity value on atomic hint patterns (NFC) (#230478)
via llvm-commits
- [llvm] [NVPTX] Preserve the load chain when custom-lowering i1 loads (PR #230497)
Alex MacLean via llvm-commits
- [llvm] [PowerPC] Handle splatted bitcast(truncate) using vbpermq (PR #227090)
via llvm-commits
- [llvm] [LV] Implement legality checks for compressing patterns (PR #230546)
Benjamin Maxwell via llvm-commits
- [llvm] [mlgo] Allow passing pre-emitc-ed models (PR #227941)
Mircea Trofin via llvm-commits
- [llvm] [SPIR-V] Type the capture_event_profiling_info value pointer as i8 (PR #230548)
Arseniy Obolenskiy via llvm-commits
- [llvm] c90b641 - [mlgo] Allow passing pre-emitc-ed models (#227941)
via llvm-commits
- [llvm] [MLGO] Do not cap evictions when logging default advisor decisions (PR #230502)
Arseniy Obolenskiy via llvm-commits
- [llvm] [SelectionDAG] Avoid rounding f32->bf16 FP_ROUND when Trunc==1 (PR #230557)
Emilio Cota via llvm-commits
- [llvm] e6cdca0 - [llvm][UnicodeData] Check for LIBXML_READER_ENABLED before building UnicodeCharSetsGenerator (#230254)
via llvm-commits
- [llvm] [SelectionDAG] Expand wide UREM by constant when there is no libcall (PR #230564)
Doha Nam via llvm-commits
- [llvm] [DAG] Implement scalarization for mask_beforefirst (PR #230556)
Paul Walker via llvm-commits
- [llvm] CodeGen: Compute live-outs of a split critical edge while updating LiveIntervals (PR #230499)
Nikita Popov via llvm-commits
- [llvm] [AMDGPU] Eliminate redundant SGPR restores (PR #222680)
Lukas Sommer via llvm-commits
- [lld] [llvm] [lld][bp] Fix page fault area calculation (PR #230566)
Ellis Hoag via llvm-commits
- [llvm] [SLP]Fix deps for stores ahead of may-throw calls (PR #230567)
Alexey Bataev via llvm-commits
- [llvm] 68f9424 - [SLP]Fix deps for stores ahead of may-throw calls
via llvm-commits
- [llvm] [LV] Fold middle-block branch in post-dissolution removeBranchOnConst. (PR #204243)
Florian Hahn via llvm-commits
- [clang] [clang-tools-extra] [lld] [lldb] [llvm] [mlir] [NFC][SystemZ][z/OS] Replace isText with OpenFlags in MemoryBuffer::getFile()/getFileOrSTDIN() (PR #228189)
via llvm-commits
- [llvm] [AMDGPU] Don't combine DS_READs that prevent WMMA latency hiding (PR #216922)
Vigneshwar Jayakumar via llvm-commits
- [llvm] [Passes] Run MergeICmps/ExpandMemMmps in LTO post-link pipeline (PR #230562)
Nikita Popov via llvm-commits
- [llvm] 3b00a39 - [NFC][AMDGPU] Add test for indexing into named barrier arrays (#230515)
via llvm-commits
- [llvm] [NFC][AMDGPU] Add test for indexing into named barrier arrays (PR #230515)
Shilei Tian via llvm-commits
- [clang] [llvm] [AMDGPU] Make named barrier type 1 byte to fix barrier IDs of array elements (PR #230516)
Shilei Tian via llvm-commits
- [llvm] [SDAG] Don't expand forceExpandWideMUL to libcall with unsuitable types. (PR #230522)
Craig Topper via llvm-commits
- [llvm] [NFC][AMDGPU] Inline `getEffectiveWavesPerEU` (PR #221600)
Shilei Tian via llvm-commits
- [llvm] [SLP] Don't drop useful cases when comparing states in isCopyablePreferable() (PR #230580)
Ryan Buchner via llvm-commits
- [llvm] [VectorCombine][TTI] Disable Load Widening for SPIR-V (PR #227623)
Gregory Roth via llvm-commits
- [llvm] 254efe1 - [AArch64] Add NVCAST to inputs of AArch64ISD::REV32/REV64 to match SDTypeProfile. (#230211)
via llvm-commits
- [llvm] [Bazel][llvm] Generate InlinerModels and RegAllocEvictModels headers (PR #230576)
Anshul Nigham via llvm-commits
- [llvm] baf56e3 - [RISCV] Correct the CFA offsets for the stack probe loop. (#230322)
via llvm-commits
- [llvm] 0258719 - [Hexagon][NFC] Fix duplicate A2_tfr check in HexagonGlobalScheduler (#227762)
via llvm-commits
- [llvm] [DebugInfo] Skip debug history clobber scans without register locations (NFC) (PR #230590)
Cullen Rhodes via llvm-commits
- [llvm] 8b087a7 - HipStdPar: Stop redirecting f64 exp and exp2 (#230573)
via llvm-commits
- [llvm] HipStdPar: Stop redirecting f64 exp and exp2 (PR #230573)
Matt Arsenault via llvm-commits
- [llvm] 8e242e6 - [LLVMABI][AARCH64] Implement Pure Scalable Type handling (#227504)
via llvm-commits
- [llvm] [ARM] Add VECTOR_REG_CAST to VBSP operands in PerformORCombine. (PR #230604)
Craig Topper via llvm-commits
- [llvm] 1be9bdf - [RISC-V][MC] Add test for C/Zce implication in .option arch (#230298)
via llvm-commits
- [llvm] [RISC-V][MC] Validate instructions prior to emission in AsmParser (PR #171739)
Alexander Richardson via llvm-commits
- [llvm] [IR][CodeGen] Add new SizedCapabilityType for use in CodeGen (PR #177192)
Alexander Richardson via llvm-commits
- [llvm] [AMDGPU] Support FLAT_SCRATCH in SILoadStoreOptimizer const offset promotion (PR #212507)
Arseniy Obolenskiy via llvm-commits
- [llvm] [NVPTX] Postprocess selection DAG to merge I16 extraction (PR #230570)
via llvm-commits
- [llvm] [VectorCombine] Add instructions created by IRBuilder into Worklist automatically (PR #230618)
Min-Yih Hsu via llvm-commits
- [llvm] da913c2 - [SeparateConstOffsetFromGEP] Track cast state during offset extraction (#229839)
via llvm-commits
- [llvm] [RISCV] Correct violations of SDTCisSameNumEltsAs. (PR #230612)
Luke Lau via llvm-commits
- [llvm] [APFloat] Add `APFloat::getConstant` for C++20 <numbers> constants (PR #227916)
Eric Ross via llvm-commits
- [llvm] [SelectionDAG] Implement SDTCisVec in SDNodeInfo::verifyNode (PR #230620)
Craig Topper via llvm-commits
- [llvm] [SeparateConstOffsetFromGEP] Accumulate offsets across expression trees (PR #230637)
Alex MacLean via llvm-commits
- [llvm] [KnownFPClass] Teach KnownFPClass that `sinh(non-positive) = non-positive` (PR #230627)
Eric Ross via llvm-commits
- [llvm] [abandoned][HLSL] Migrate pre-commit HLSL CI to Windows (and off Beanz's closet Mac) (PR #226312)
Helena Kotas via llvm-commits
- [llvm] [GlobalISel] Port simplifyDemandedBits for G_AND and G_OR (PR #198808)
Alan Li via llvm-commits
- [llvm] [GlobalISel] Port `SimplifyDemandedBits` over (PR #198808)
Alan Li via llvm-commits
- [llvm] CodeGen: Find PHI sources while rewriting PHIs in SplitCriticalEdge (PR #230500)
Matt Arsenault via llvm-commits
- [llvm] 4b9d60e - CodeGen: Find PHI sources while rewriting PHIs in SplitCriticalEdge (#230500)
via llvm-commits
- [llvm] [HLSL] Migrate pre-commit HLSL CI to Windows (and off Beanz's closet Mac) (PR #230639)
Helena Kotas via llvm-commits
- [llvm] ba8acdd - PowerPC: Mark TLS call pseudo clobbers dead at build time (#227050)
via llvm-commits
- [llvm] [LV] Add tests for FindLast reductions with nested phis (NFC). (PR #230640)
Florian Hahn via llvm-commits
- [llvm] 156fd40 - [SCEV] Scale canReuseInstruction's walk budget with the expression size. (#230508)
via llvm-commits
- [llvm] [InlineCost][Draft] Add discount for clobbered loads (PR #230641)
Ethan Luis McDonough via llvm-commits
- [llvm] [SelectionDAG][NVPTX] Stop casting atomic xchg to integer in AtomicExpand; rely on legalization (PR #229867)
Yonah Goldberg via llvm-commits
- [llvm] [SLP][NFC]Add a test with incorrect blended loads alignment, NFC (PR #230642)
Alexey Bataev via llvm-commits
- [llvm] [VPlan] Allow differing IV types in strict multi-use-reductions (PR #229759)
Ramkumar Ramachandra via llvm-commits
- [llvm] [Offload][OpenMP] Report source location of data transfers and synchr… (PR #223730)
Jason Van Beusekom via llvm-commits
- [llvm] 75b7149 - [SLP]Fix gather of extractelements from different-sized vectors (#230664)
via llvm-commits
- [llvm] [SLP]Fix gather of extractelements from different-sized vectors (PR #230664)
Alexey Bataev via llvm-commits
- [lld] [llvm] [llvm] Refactor `isLargeGlobalValue` from `TargetMachine` to `GlobalValue` (PR #230666)
Arthur Eubanks via llvm-commits
- [llvm] 8e40b79 - AMDGPU: Stop relying on -amdgpu-scalarize-global-loads=false in ALU tests (#229924)
via llvm-commits
- [llvm] AMDGPU: Index loads by workitem id in tests shared with r600 (PR #230667)
Matt Arsenault via llvm-commits
- [llvm] [RISCV] Reassociate add (add X, (ext Y)), (ext Z) -> add (add (ext Y), (ext Z)), X (PR #230657)
Craig Topper via llvm-commits
- [llvm] [RISCV] Reassociate add (add X, (ext Y)), (ext Z) -> add (add (ext Y), (ext Z)), X (PR #230657)
Luke Lau via llvm-commits
- [llvm] [RISCV] Reassociate add (add X, (ext Y)), (ext Z) -> add (add (ext Y), (ext Z)), X (PR #230657)
Luke Lau via llvm-commits
- [llvm] [RISCV] Reassociate add (add X, (ext Y)), (ext Z) -> add (add (ext Y), (ext Z)), X (PR #230657)
Luke Lau via llvm-commits
- [llvm] [AMDGPU] Add async and tensor waits to `isWaitcnt` (PR #230663)
Krzysztof Drewniak via llvm-commits
- [llvm] [Offload][Omp] Ensure errors are consumed when debug is off (PR #230668)
Alex Duran via llvm-commits
- [llvm] [SLP][NFC]Add a test with incorrect nneg propagation, NFC (PR #230670)
Alexey Bataev via llvm-commits
- [llvm] [SelectionDAG] Implement SDTCisSubVecOfVec in SDNodeInfo::verifyNode. (PR #230628)
Sergei Barannikov via llvm-commits
- [llvm] [RISCV] Make rv64 i64 -> i32 trunc free (PR #230681)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SelectionDAG] Verify SDTCisSameSizeAs in SDNodeInfo::verifyNode (PR #230634)
Sergei Barannikov via llvm-commits
- [llvm] [SelectionDAG] Verify SDTCisSameNumEltsAs in SDNodeInfo::verifyNode (PR #230616)
Luke Lau via llvm-commits
- [clang] [llvm] [Clang][C++23] Core language changes from P1467R9 extended floating-point types and standard names. (PR #78503)
Oleksandr Kulkov via llvm-commits
- [llvm] [llvm][GlobalOpt] Preserve COMDATs for Dead GVs (PR #222720)
Eli Friedman via llvm-commits
- [llvm] Reapply "workflows/precommit: Use single shared job to compute project" (PR #230203)
Tom Stellard via llvm-commits
- [lld] [llvm] [nfc][llvm] Refactor `isLargeGlobalValue` from `TargetMachine` to `GlobalValue` (PR #230666)
Alan Zhao via llvm-commits
- [clang] [llvm] [Clang] Support `--offload-compress` for offload binaries (PR #227765)
Joseph Huber via llvm-commits
- [llvm] [SLP]Fix crash for widened strided loads/stores of pointers (PR #230683)
Alexey Bataev via llvm-commits
- [llvm] da1afa7 - [SLP]Fix crash for widened strided loads/stores of pointers
via llvm-commits
- [llvm] [AMDGPU] Add SSA-form memory clause pass (AMDGPUFormSSAMemoryClauses) (PR #209656)
Jun Wang via llvm-commits
- [llvm] 48026d4 - [RISCV] Custom lower fixed length mask_beforefirst (#230482)
via llvm-commits
- [compiler-rt] fix: add missing define in asan_mapping_sparc64.h (PR #230686)
Andrew Haberlandt via llvm-commits
- [llvm] [SelectionDAG] Make `AND`/`OR `demanded bit simplification symmetric (PR #230680)
Alan Li via llvm-commits
- [llvm] [cmake] Pass arguments to ccache on Windows (PR #218791)
via llvm-commits
- [llvm] [IR][SelectionDAG] Add `llvm.vector.shuffle` intrinsic (Part 1). (PR #219259)
Oscar Smith via llvm-commits
- [llvm] [IR][SelectionDAG] Add `llvm.vector.shuffle` intrinsic (Part 1) (PR #219259)
Oscar Smith via llvm-commits
- [lld] [Mach-O] Parallelize ICF's section sort. NFC (PR #230269)
Nico Weber via llvm-commits
- [lld] b0fce46 - [Mach-O] Parallelize ICF's section sort. NFC (#230269)
via llvm-commits
- [llvm] [X86] Keep upper-16-bit f32 extractions in XMM for f16/v8i16 (PR #230698)
Emilio Cota via llvm-commits
- [llvm] HipStdPar: Don't reprocess math library functions as intrinsics (PR #230603)
Dmitry Sidorov via llvm-commits
- [llvm] [SelectionDAG] Make AND/OR demanded-bit simplification symmetric (PR #230700)
Alan Li via llvm-commits
- [llvm] [llvm-readobj][ELF] Drop redundant counts from --call-graph-info output (PR #230647)
Prabhu Rajasekaran via llvm-commits
- [llvm] 5afa2c0 - [RISCV] Reassociate add (add X, (ext Y)), (ext Z) -> add (add (ext Y), (ext Z)), X (#230657)
via llvm-commits
- [llvm] [llvm-gsymutil] Add --stubs-file to name mach-o stubs from file (PR #230285)
Greg Clayton via llvm-commits
- [llvm] [X86] Do not lower f16 to i128 conversions as i64 on Win64 (PR #230712)
Aochang Liu via llvm-commits
- [llvm] [LV] Use speculative load intrinsics for early-exit vectorization (PR #180039)
Ramkumar Ramachandra via llvm-commits
- [llvm] [SCEV] Use signed form of samesign compares when it tightens exit limits (PR #223514)
Hao Ren via llvm-commits
- [llvm] [Attributor] Skip the dead-internal-function walk for AAs that will not be updated (PR #227187)
Farid Zakaria via llvm-commits
- [llvm] [Scalarizer] Support selecting which opcodes to scalarize (PR #230715)
via llvm-commits
- [llvm] [CodeGen][AArch64][X86] Add target-independent conditional-compare formation pass (PR #219359)
Feng Zou via llvm-commits
- [clang] [compiler-rt] [compiler-rt:ubsan] Add -fsanitize=unaligned-pointer-subtraction to detect undefined subtraction of two pointers (PR #210635)
Jerry Dang via llvm-commits
- [llvm] [AMDGPU] DAG Mutation to solve load bunching in outer product matrix multiplications (PR #203095)
Axel Sorenson via llvm-commits
- [llvm] AMDGPU: Remove remaining uses of -amdgpu-scalarize-global-loads=false (PR #230198)
John Hui via llvm-commits
- [llvm] [AArch64LoadStoreOptimizer] Pair far-offset loads/stores via base-register adjustment (PR #223684)
via llvm-commits
- [llvm] [RISCV][GlobalISel] Use Zfhmin for G_FPEXT from f16 to f64 (PR #230721)
Zeyi Xu via llvm-commits
- [llvm] Debug llvm pahole (PR #230615)
via llvm-commits
- [llvm] [Bazel] Fixes fa70420 (PR #230722)
via llvm-commits
- [llvm] [RISCV][GlobalISel] Use Zfhmin for G_FCONSTANT s16 (PR #230723)
Zeyi Xu via llvm-commits
- [llvm] [MLGO] Remove std includes from pre-generated test models (PR #230724)
John Hui via llvm-commits
- [llvm] [BOLT] Fix address map parsing for 32-bit targets (PR #229567)
via llvm-commits
- [llvm] [X86] Fold compares of widened vectors into masked compares (PR #228350)
Phoebe Wang via llvm-commits
- [llvm] Hidden vis ci (PR #230727)
Tom Stellard via llvm-commits
- [llvm] [SupportTests] Ignore STATUS_DLL_NOT_FOUND errors on Windows (PR #230726)
Saleem Abdulrasool via llvm-commits
- [llvm] [BOLT] Restore -relax-plt as a deprecated no-op option (PR #230729)
Rafael Auler via llvm-commits
- [llvm] [AsmPrinter] Record direct calls through the GOT in the call graph section (PR #230730)
Prabhu Rajasekaran via llvm-commits
- [llvm] [Mips] Remove unused -mips-fix-global-base-reg (PR #230733)
Fangrui Song via llvm-commits
- [llvm] 006de98 - [Mips] Remove unused -mips-fix-global-base-reg (#230733)
via llvm-commits
- [llvm] d19748b - [MLGO] Remove std includes from pre-generated test models (#230724)
via llvm-commits
- [llvm] [AArch64] Allow commuted `min` operands in clamp patterns (PR #230693)
Usman Nadeem via llvm-commits
- [llvm] [InlineSpiller][AMDGPU] Skip undef lanes when spilling a partial-def super-register (PR #226689)
Quentin Colombet via llvm-commits
- [llvm] [Xtensa] Route windowed-ABI stack realignment through MOVSP (PR #230735)
via llvm-commits
- [llvm] [flang-rt] Fix FORT_CONVERT_UNIT parsing hang when strtok_r nulls the save pointer (PR #230736)
Kelvin Li via llvm-commits
- [llvm] [Mips] Declare command line options in TableGen (PR #230738)
Fangrui Song via llvm-commits
- [compiler-rt] [compiler-rt][ARM] Mark arm/adddf3.S as superseding arm/aeabi_drsub.c (PR #230739)
Sharjeel Khan via llvm-commits
- [llvm] [llvm] Fix enum-enum bitwise and arithmetic conversions for C++26 (PR #228945)
via llvm-commits
- [llvm] [ControlFlowHub] Preserve scaled branch weights on ControlFlowHub guard branches (PR #230740)
Alok Kumar Sharma via llvm-commits
- [clang] [llvm] [Instrumentation] Stop exporting cl::opts. NFC (PR #230725)
Alexis Engelke via llvm-commits
- [llvm] [KnownFPClass] Refine positive zero result for sqrt (PR #214987)
Eric Ross via llvm-commits
- [llvm] Revert "[KnownFPClass] Refine positive zero result for sqrt" (PR #230741)
Eric Ross via llvm-commits
- [llvm] 53310c9 - [Instrumentation] Stop exporting cl::opts. NFC (#230725)
via llvm-commits
- [llvm] ff3ba2b - Revert "[KnownFPClass] Refine positive zero result for sqrt" (#230741)
via llvm-commits
- [llvm] [Bazel] Fixes 53310c9 (PR #230743)
via llvm-commits
- [llvm] AMDGPU: Remove amdgpu-scalarize-global-loads option (PR #230744)
Matt Arsenault via llvm-commits
- [llvm] [ASan] Remove unused -asan-debug (PR #230745)
Fangrui Song via llvm-commits
- [llvm] 2f14ed9 - [ASan] Remove unused -asan-debug (#230745)
via llvm-commits
- [llvm] [Instrumentation] Declare command line options in TableGen (PR #230746)
Fangrui Song via llvm-commits
- [llvm] [LFI][AArch64] Account for deferred LR guard flush in getInstSizeInBytes (PR #224791)
Sharjeel Khan via llvm-commits
- [llvm] [X86] Add tuning for fast AVX2 vector division (PR #219873)
Xiaomeng Zhang via llvm-commits
- [llvm] [CodeGen] Place unsplittable blocks after the profiled blocks in BB sections (PR #228173)
LLVM Continuous Integration via llvm-commits
- [llvm] [Analysis][CodeGen] Remove unused haveFastClmul (NFC) (PR #230748)
Kazu Hirata via llvm-commits
- [llvm] [Mips] Use CompressPat for microMIPS instruction compression (PR #224217)
Ying Huang via llvm-commits
- [llvm] [Coroutines] Use SmallMapVector for AllocaInfo::Aliases (#230585) (PR #230749)
via llvm-commits
- [llvm] [orc-rt] Simplify SimpleSymbolTable::makeIncompatibleDefsError (PR #230750)
Lang Hames via llvm-commits
- [llvm] f7e2b83 - AMDGPU: Remove amdgpu-scalarize-global-loads option (#230744)
via llvm-commits
- [clang] [lld] [llvm] [Clang][WebAssembly] Add wasm_global declaration attribute (PR #221909)
Changqing Jing via llvm-commits
- [llvm] [ARM] Implement CRC32 const folding (PR #230751)
Shreesh Adiga via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Nikita Popov via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Nikita Popov via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Nikita Popov via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Andreas Jonson via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Nikita Popov via llvm-commits
- [clang] [llvm] [InstCombine] Generalizing icmp eq/ne X, 0/1 to trunc X to i1 IIF X in range (0,2] (PR #184182)
Andreas Jonson via llvm-commits
- [llvm] [IR] Return poison instead of undef for poison shufflevector elements (PR #225598)
Le Gia Bao via llvm-commits
- [llvm] Misched unify sunit print format (PR #229756)
Tomer Shafir via llvm-commits
- [llvm] [MISched] Unify SUnit formatting among users (PR #229756)
Tomer Shafir via llvm-commits
- [clang] [flang] [llvm] Initial support for vectorization using HVML (Hygon vector math library) (PR #219098)
Xiaomeng Zhang via llvm-commits
- [llvm] [Coverage] Clamp negative counter expression values to zero (PR #230752)
Snehal Reddy via llvm-commits
- [llvm] [orc-rt] Use const void* for SymbolLookupResults. (PR #230753)
Lang Hames via llvm-commits
- [llvm] 8e24f8f - [Coroutines] Use SmallMapVector for AllocaInfo::Aliases (#230585) (#230749)
via llvm-commits
- [llvm] b16fe31 - [Analysis][CodeGen] Remove unused haveFastClmul (NFC) (#230748)
via llvm-commits
- [llvm] [Attributor] Don't derive known memory behavior of floating values from the instruction itself (PR #230755)
Divyansh Yadav via llvm-commits
- [llvm] [VPlan] Strip call to isTruncateFree from narrowIndTrunc (PR #230756)
Ramkumar Ramachandra via llvm-commits
- [llvm] Fix Attributor crash on mismatched callee return types (PR #230757)
Changqing Jing via llvm-commits
- [llvm] [PowerPC] Custom-lower `f128` `BR_CC` for VSX w/o P9 (PR #230758)
Alex Rønne Petersen via llvm-commits
- [llvm] [Passes] Honor module Attributor options in FullLTO (PR #230759)
Changqing Jing via llvm-commits
- [lld] [lld][ELF] Use explicit size_t cast in SyntheticSections (PR #230760)
QIU Chaofan via llvm-commits
- [llvm] [orc-rt] Add SymbolLookupSet-based lookup to SimpleSymbolTable (PR #230761)
Lang Hames via llvm-commits
- [llvm] [GlobalOpt] Fix alignments of globals introduced for allocations (PR #216480)
Ömer Sinan Ağacan via llvm-commits
- [llvm] [VPlan] Model the connected main-loop CFG and epilogue resume values. (PR #229921)
Ramkumar Ramachandra via llvm-commits
- [llvm] AMDGPU: Restore kernel tests in early-if-convert-cost.ll (PR #230764)
Matt Arsenault via llvm-commits
- [llvm] [AMDGPU] Allow convert SALU fptrunc.round to VALU (PR #230533)
Russell Liu via llvm-commits
- [llvm] [InstCombine] Fold eq/ne between scmp(X, 0) and zext(V != 0) (PR #230767)
via llvm-commits
- [llvm] 0aa8163 - [AMDGPU] Don't select scalar loads for LDS/scratch (#227423)
via llvm-commits
- [llvm] b2228ad - AMDGPU: Restore kernel tests in early-if-convert-cost.ll (#230764)
via llvm-commits
- [llvm] [RISCV] Optimize CLMULH of i32 loads on RV64 (PR #228669)
via llvm-commits
- [llvm] [RISCV] Optimize repeated-half i64 vector splats on RV32 (PR #228119)
via llvm-commits
- [llvm] [Object] Avoid integer overflow in big archive symbol table bounds checks (PR #230771)
Prasad Deshpande via llvm-commits
- [llvm] [AArch64] Remove tryShiftAmountMod - gap finding (PR #228753)
Deepak Shirke via llvm-commits
- [llvm] [VPlan] Re-use LCSSA phis when expanding AddRecs of other loops. (PR #230772)
Florian Hahn via llvm-commits
- [llvm] [InstCombine] Fold unsigned icmp of nuw adds with constants (PR #226685)
Karl-Johan Karlsson via llvm-commits
- [llvm] [LAA] Add tests for a set of miscompiles. (NFC) (PR #230773)
Florian Hahn via llvm-commits
- [llvm] [APFloat] Correct upper boundary distances for directed decimal rounding (PR #230774)
via llvm-commits
- [llvm] [libc][ci] Update FreeBSD version to 15 and bump VM action (PR #230775)
Connector Switch via llvm-commits
- [llvm] [Object] Validate the big archive global symbol table's symbol count (PR #230777)
Prasad Deshpande via llvm-commits
- [llvm] [SelectionDAG] Convert f16 to wide integers without a libcall (PR #230779)
Aochang Liu via llvm-commits
- [llvm] RuntimeLibcalls: Remove the Default* libcall lists (PR #230405)
Matt Arsenault via llvm-commits
- [llvm] b0e32db - [LAA] Use EarliestEscapeAnalysis for alias queries. (#228179)
via llvm-commits
- [llvm] [Object] Bounds-check the big archive next member offset (PR #230781)
Prasad Deshpande via llvm-commits
- [llvm] bbcd4d5 - [VPlan] Re-use LCSSA phis when expanding AddRecs of other loops. (#230772)
via llvm-commits
- [llvm] 26183f3 - [LAA] Add tests for a set of miscompiles. (NFC) (#230773)
via llvm-commits
- [llvm] [IR] Verify DISubrange/DISubrangeType bounds are ConstantInt (PR #223134)
Sewon Ahn via llvm-commits
- [lld] [lld][ELF] Refresh .eh_frame_hdr cache after cfi jump table relaxation (PR #228218)
Martin Levesque via llvm-commits
- [llvm] [DAGCombiner] Narrow the integer source of uint_to_fp (PR #222899)
Dmitry Sidorov via llvm-commits
- [llvm] [DAGCombiner] Narrow the integer source of sint_to_fp (PR #224242)
Dmitry Sidorov via llvm-commits
- [llvm] [X86] Expand vector atomicrmw through a cmpxchg loop (PR #230788)
Krishnan R via llvm-commits
- [llvm] [SimplifyCFG] Merge identical landing pads when the successor has phis (PR #230789)
Timur Golubovich via llvm-commits
- [llvm] [InstCombine] fold add (add x, 1), (or (add x, 1), -2) -> and x, -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold add (add x, 1), (or (add x, 1), -2) -> and x, -2 (PR #230791)
via llvm-commits
- [llvm] [InstCombine] fold add (add x, 1), (or (add x, 1), -2) -> and x, -2 (PR #230791)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold add (add x, 1), (or (add x, 1), -2) -> and x, -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold add (or x, -2), x -> and (add x, -1), -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold (A | -2) + A -> (A - 1) & -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold (A | -2) + A -> (A - 1) & -2 (PR #230791)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold (A | -2) + A -> (A - 1) & -2 (PR #230791)
Andreas Jonson via llvm-commits
- [clang] [llvm] [Xtensa][clang] Add Intel Audio DSP (ADSP) processor definitions (PR #228927)
Liam Girdwood via llvm-commits
- [llvm] [X86] Avoid x87-dependent FP-to-i64 lowering without x87 (PR #230792)
via llvm-commits
- [clang] [lld] [llvm] [clang][lld] Apple Kext support (PR #204805)
Visual Ehrmanntraut via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold sub (and (add x, pow2), ~pow2), x -> and x, pow2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Nikita Popov via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (PR #230796)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] Optimize and of overflow checks (PR #141962)
Bobby SONG via llvm-commits
- [llvm] [ThinLTO] Optionally import noinline callees at critical call sites (PR #229540)
Teresa Johnson via llvm-commits
- [compiler-rt] 762316e - [compiler-rt][ROCm] Collect Linux profiles from resident images (#229518)
via llvm-commits
- [llvm] [memprof] Remove MemProf format Version 2 (PR #229262)
Teresa Johnson via llvm-commits
- [llvm] [LoopVectorize] Introduce check-first vectorization for multiple early-exit loops (PR #210492)
Aditya Trivedi via llvm-commits
- [llvm] [LoongArch] Implement computeKnownBitsForTargetNode() cases for `V{ALL,ANY}_{,NON}ZERO` (PR #230799)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Select branches on `V{ALL,ANY}_{,NON}ZERO` to consume $FCC directly (PR #230801)
Rong Mantle Bao via llvm-commits
- [llvm] [LoongArch] Materialize vector zero-test results via `MOVCF2GR` (PR #230804)
Rong Mantle Bao via llvm-commits
- [lld] [lld][WebAssembly] Handle locally bound relative data relocations in PIC (PR #230805)
Anutosh Bhat via llvm-commits
- [llvm] [X86] Ignore debug-only blocks in LVI gadget graph traversal (PR #225994)
Demetrios Chiuratto Agourakis via llvm-commits
- [llvm] [IRCE] Eliminate inverted range checks (PR #230222)
via llvm-commits
- [llvm] [KnownFPClass] Propagate sign information through atan (PR #229902)
Youssef W. Abdallah via llvm-commits
- [clang] [lld] [llvm] [lld-macho][clang][llvm-objcopy] Add x86_64 kext bundle support (PR #204805)
Visual Ehrmanntraut via llvm-commits
- [llvm] [llubi] Don't expose provenance of freed memory object (PR #230810)
Yingwei Zheng via llvm-commits
- [llvm] [AArch64] Remove tryShiftAmountMod in favor of SelectShiftMask (PR #228753)
Deepak Shirke via llvm-commits
- [llvm] f621e76 - [llubi] Make sure the allocated address can be represented (#230656)
via llvm-commits
- [llvm] [llubi] Make sure the allocated address can be represented (PR #230656)
Yingwei Zheng via llvm-commits
- [llvm] 2e48c55 - [memprof] Remove MemProf format Version 2 (#229262)
via llvm-commits
- [llvm] [GlobalISel] Preserve demanded lanes in G_CTLS KnownBits queries (PR #230096)
Matt Arsenault via llvm-commits
- [llvm] e7305f7 - [GlobalISel] Preserve demanded lanes in G_CTLS KnownBits queries (#230096)
via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Alexis Engelke via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Alexis Engelke via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Alexis Engelke via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Alexis Engelke via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Fangrui Song via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
Alexis Engelke via llvm-commits
- [llvm] [Option] Add OptionalValueField, OptionalEnumField, and list fields (PR #230812)
LLVM Continuous Integration via llvm-commits
- [llvm] 5e11963 - [RuntimeDyld] Implement more relocations for 32-bit PowerPC (#229933)
via llvm-commits
- [llvm] [RuntimeDyld] Implement more relocations for 32-bit PowerPC (PR #229933)
Jiaxun Yang via llvm-commits
- [llvm] [InstCombine] fix power of two examples (PR #230814)
Andreas Jonson via llvm-commits
- [llvm] ad7bb9b - [InstCombine] fold (A | -2) + A -> (A - 1) & -2 (#230791)
via llvm-commits
- [llvm] 9797e39 - [Option] Add OptionalValueField, OptionalEnumField, and list fields (#230812)
via llvm-commits
- [llvm] [SLP][NFC]Add a test with regressed vectorization, NFC (PR #230817)
Alexey Bataev via llvm-commits
- [llvm] f5cee7d - [SLP][NFC]Add a test with regressed vectorization, NFC
via llvm-commits
- [llvm] [llubi] Fix wrong assert when writing `poison` at an unaligned bit offset (PR #230819)
Pedro Lobo via llvm-commits
- [llvm] [bazel] Port 3d519f5eee97 (PR #230820)
Benjamin Kramer via llvm-commits
- [llvm] [SLP]Count splats as single broadcast for in-loop trees (PR #230821)
Alexey Bataev via llvm-commits
- [llvm] 82e2575 - [InstCombine] fix power of two examples (#230814)
via llvm-commits
- [llvm] Infer nonzero on false OR edges (PR #230822)
via llvm-commits
- [llvm] 459bce1 - [ConstantFolding] Fold `bitinsert` and `bitextract` (#229115)
via llvm-commits
- [llvm] 46cf0b8 - [SLP]Count splats as single broadcast for in-loop trees
via llvm-commits
- [llvm] [SLP]Fix crash on split reduction root cast cost (PR #230597)
Nikita Popov via llvm-commits
- [llvm] [LoongArch] Add initial MachineCombiner support (PR #230830)
via llvm-commits
- [llvm] [SLP][NFC]Add a test missing incorrect nsw propagation on constexprs, NFC (PR #230834)
Alexey Bataev via llvm-commits
- [llvm] a35842d - [SLP][NFC]Add a test missing incorrect nsw propagation on constexprs, NFC
via llvm-commits
- [llvm] [SLP]Drop wrap flags for negated lanes with constant expressions (PR #230835)
Alexey Bataev via llvm-commits
- [llvm] deb983a - [SLP]Drop wrap flags for negated lanes with constant expressions
via llvm-commits
- [llvm] [SLP]Fix crash on gather reduction root cast cost (PR #230836)
Alexey Bataev via llvm-commits
- [llvm] d5178df - [SLP]Fix crash on gather reduction root cast cost
via llvm-commits
- [llvm] [AArch64][MachineOutliner] Allow calls to be outlined when PAC is enabled (PR #230840)
Mazen Hatem via llvm-commits
- [llvm] [llubi] Check UB before handling ret attributes (PR #230842)
Yingwei Zheng via llvm-commits
- [llvm] [BasicAA] Reuse object size when both pointers share an object (NFC) (PR #230844)
Florian Hahn via llvm-commits
- [llvm] [llubi] Correctly render large addresses in `verifyMemAccess` (PR #230845)
Yingwei Zheng via llvm-commits
- [llvm] [SLP]Fix assert on gather of out-of-bounds extract (PR #230847)
Alexey Bataev via llvm-commits
- [llvm] b1dd93c - [SLP]Fix assert on gather of out-of-bounds extract
via llvm-commits
- [llvm] [llubi] Reset retval for noop inline asm (PR #230849)
Yingwei Zheng via llvm-commits
- [llvm] [llubi] Use correct tag bitwidth to recover provenances (PR #230852)
Yingwei Zheng via llvm-commits
- [llvm] [RISC-V] Extend support for -fsanitize=shadow-call-stack (PR #230853)
Malhar Badve via llvm-commits
- [llvm] [libc][Bazel] Add missing float MPFR math tests (PR #227081)
via llvm-commits
- [llvm] [VPlan] Remove unused recipe constructor overloads (NFC) (PR #230857)
Florian Hahn via llvm-commits
- [llvm] [VPlan] Expand AddRecs in VPlan, only canonical ARs in IR. (PR #230858)
Florian Hahn via llvm-commits
- [llvm] [SCEV] Remove unused code from ScalarEvolutionExpressions.h (NFC) (PR #230859)
Florian Hahn via llvm-commits
- [llvm] [LAA] Compute MaxBTC * MaxStride in a wide enough type. (PR #230860)
Florian Hahn via llvm-commits
- [llvm] [RuntimeDyld] Fix layout-dependent REL16 checks in `ppc32_elf_reloc.s` (PR #230861)
Pedro Lobo via llvm-commits
- [llvm] [LoopIdiom] Fix and improve undef handling in store chains that become memsets (PR #230863)
Ömer Sinan Ağacan via llvm-commits
- [llvm] [PowerPC] Fix Impossible reg-to-reg copy when inline asm uses crN registers as inputs (PR #196182)
via llvm-commits
- [clang] [llvm] Repeat Clang Interpreter lit tests 100 times (PR #230864)
Vitaly Buka via llvm-commits
- [llvm] 444a376 - [RISCV][P-ext] Combine PPAIR isel patterns by using HwMode types. NFC (#230357)
via llvm-commits
- [llvm] 1a54676 - [RuntimeDyld] Fix layout-dependent REL16 checks in `ppc32_elf_reloc.s` (#230861)
via llvm-commits
- [llvm] [DebugInfo][DwarfDebug] Defer parent attachment of function-local DIEs until endModule() (PR #230866)
Vladislav Dzhidzhoev via llvm-commits
- [llvm] [DebugInfo][DwarfDebug] Scan for inlined subprograms in LexicalScopes::initialize() (PR #226536)
Vladislav Dzhidzhoev via llvm-commits
- [llvm] 86d42a0 - [DAGCombiner] Narrow the integer source of uint_to_fp (#222899)
via llvm-commits
- [llvm] dc3cff3 - VE: Remove broken nested call frame around dynamic stack allocation (#229011)
via llvm-commits
- [llvm] [CMake] Special-case HarmonyOS hosts in get_host_triple() (PR #219882)
via llvm-commits
- [libc] [llvm] [libc][math][c23] Fix signed zeros in `tanpif` and `tanpif16` (PR #230371)
Connector Switch via llvm-commits
- [llvm] [gn] "port" 5a1dbdd2b16d3 (PR #230871)
Nico Weber via llvm-commits
- [llvm] [gn build] Port 8c4c52bb62ee (PR #230872)
Nico Weber via llvm-commits
- [llvm] [gn build] Port a09df97a747bb (PR #230873)
Nico Weber via llvm-commits
- [llvm] [gn build] Port dba1b67f855fb (PR #230874)
Nico Weber via llvm-commits
- [llvm] a9fd78b - [gn] "port" 5a1dbdd2b16d3 (#230871)
via llvm-commits
- [llvm] daf014f - [gn build] Port 8c4c52bb62ee (#230872)
via llvm-commits
- [llvm] 16c05cd - [gn build] Port dba1b67f855fb (#230874)
via llvm-commits
- [llvm] 2d194c6 - [gn build] Port a09df97a747bb (#230873)
via llvm-commits
- [llvm] [orc-rt] Add SPSSymbolLookupSet typedef, clean up users. (PR #230876)
Lang Hames via llvm-commits
- [llvm] 611eee0 - [SLP]Vectorize consecutive loads with undef lanes as a wide load
via llvm-commits
- [llvm] [orc-rt] Add SPSSymbolLookupResult typedef, clean up users. (PR #230877)
Lang Hames via llvm-commits
- [llvm] 024cc93 - [AMDGPU] Allow commuting immediates out of src0 when legal (#229906)
via llvm-commits
- [llvm] cbdad67 - [AMDGPU] Form VOPD dot2 pairs with a literal in src1 (#230183)
via llvm-commits
- [llvm] [NFC] Fix wrong comment in insertFastDivAndRem (PR #230879)
via llvm-commits
- [clang] [llvm] [ORC] Use MapperJITLinkMemoryManager to avoid "out of range of Delta32 fixup" (PR #203133)
Vitaly Buka via llvm-commits
- [llvm] [orc-rt] Default LockedAccess's LockT arg to std::scoped_lock (PR #230881)
Lang Hames via llvm-commits
- [llvm] d283ad8 - [NVPTX] Preserve the load chain when custom-lowering i1 loads (#230497)
via llvm-commits
- [llvm] [SimplifyLibCalls] Move constant folding logic in `ConstantFoldLibCall2` (PR #172139)
Ron Xavier via llvm-commits
- [llvm] [ProfileData] Move ProfCorrelatorKind out of InstrProfCorrelator (PR #230883)
Fangrui Song via llvm-commits
- [llvm] [AMDGPU] Use 8-bit barrier member count on GFX13 (PR #230884)
Chinmay Deshpande via llvm-commits
- [llvm] [X86] promote f16 `frexp` (PR #208530)
via llvm-commits
- [llvm] [BypassSlowDivision] Add option to bypass div/rem with constant divisors (PR #230885)
via llvm-commits
- [clang] [llvm] [ThinLTO][SplitModuleCG] Add call-graph based module splitting (1/N) (PR #198702)
Shilei Tian via llvm-commits
- [llvm] [SelectionDAG] Soften STRICT_FP16_TO_FP and STRICT_BF16_TO_FP (PR #230889)
Aochang Liu via llvm-commits
- [llvm] [InstCombine] Fold !umul_ov(X, C1) && X*C1 <u C2 to X <u ceil(C2/C1) (PR #227293)
via llvm-commits
- [llvm] [LSR] Return null from `getExactSDiv` for a zero divisor (PR #226703)
Akash Manna via llvm-commits
- [llvm] [Option] Add Condition for build-conditional library options (PR #230898)
Fangrui Song via llvm-commits
- [llvm] 1abc2f0 - [AMDGPU] Use 8-bit barrier member count on GFX13 (#230884)
via llvm-commits
- [llvm] [AMDGPU] Add and_or tests for literals shared across ThreeOp pairs (NFC) (PR #224198)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU] Avoid fract fold when result is observable at +/-inf (PR #219515)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AMDGPU] Relax constant bus limit for widely shared constants (PR #224049)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AArch64] Ignore debug labels when folding settag SP updates (PR #230904)
Yongqiang Tian via llvm-commits
- [llvm] d05425e - [ProfileData] Move ProfCorrelatorKind out of InstrProfCorrelator (#230883)
via llvm-commits
- [llvm] CodeGen: Initialize TargetLoweringObjectFile from MachineModuleInfo (PR #226836)
Fangrui Song via llvm-commits
- [llvm] a73ed0c - [Support] Fix SpecificBumpPtrAllocator move assignment leak (#230480)
via llvm-commits
- [llvm] [Analysis][ARM] Remove unused getNumBytesToPadGlobalArray (NFC) (PR #230912)
Kazu Hirata via llvm-commits
- [llvm] 5224660 - [ARM] Implement CRC32 const folding (#230751)
via llvm-commits
- [llvm] [Verifier][IR] Make sure bitcast doesn't change type size (PR #230914)
Yingwei Zheng via llvm-commits
- [llvm] AMDGPU: Don't use PressureDiffs with implicit allocatable physregs (PR #230916)
Matt Arsenault via llvm-commits
- [llvm] LV: Expand llvm.histogram intrinsic to support umax, umin, and uadd.sat operations (PR #127399)
via llvm-commits
- [llvm] 0c4b521 - [AMDGPU] Insert wmma coexec aware anti-hints rules (#226397)
via llvm-commits
- [llvm] 7d19366 - [llubi] Correctly render large addresses (#230845)
via llvm-commits
- [llvm] 118b80a - [llubi] Check UB before handling ret attributes (#230842)
via llvm-commits
- [clang] [llvm] [IR] Return poison instead of undef for poison shufflevector elements (PR #225598)
Le Gia Bao via llvm-commits
- [llvm] [ValueTracking] Add trunc support in isKnownNonZeroFromAssume (PR #230919)
Andreas Jonson via llvm-commits
- [clang] [lld] [llvm] [Alignment] Add a static function to construct an `Align` from power-of-2 values (PR #230920)
Ömer Sinan Ağacan via llvm-commits
- [clang] [llvm] [SPIR-V] Fix enqueue_kernel crash on clang -O0 local sizes (PR #230924)
Arseniy Obolenskiy via llvm-commits
- [llvm] [ValueTracking] Handle trunc to i1 condition in isKnownNonNullFromDominatingCondition (PR #230925)
Andreas Jonson via llvm-commits
- [llvm] [InstCombine] Remove redundant CRC32 intrinsic declarations from tests (PR #230926)
Shreesh Adiga via llvm-commits
- [llvm] 25a8bb5 - [RISCV] Fix FP properties for Zvbdota instructions (#228737)
via llvm-commits
- [llvm] [RISCV] Fix FP properties for Zvbdota instructions (PR #228737)
Zeyi Xu via llvm-commits
- [llvm] [InstCombine] Lower ilogb to bit operations for normal values (PR #218044)
via llvm-commits
- [llvm] [WIP][RISCV] Reassociate sext's in sext/add chain (PR #230928)
Ryan Buchner via llvm-commits
- [llvm] cb23a36 - CodeGen: Initialize TargetLoweringObjectFile from MachineModuleInfo (#226836)
via llvm-commits
- [llvm] CodeGen: Merge TargetLoweringObjectFile::getModuleMetadata into Initialize (PR #226837)
Matt Arsenault via llvm-commits
- [llvm] [AVR] Support STDSPQRr on reduced tiny cores (PR #230226)
Ben Shi via llvm-commits
- [llvm] ceae108 - [GlobalISel] [AArch64] Support ROTR shift amount masking in SelectShiftMask (#226579)
via llvm-commits
- [llvm] CodeGen: Remove stale last-block case when splitting a critical edge (PR #230930)
Matt Arsenault via llvm-commits
- [llvm] TwoAddressInstructions: Keep undef INSERT_SUBREG lanes defined (PR #230931)
Matt Arsenault via llvm-commits
- [llvm] 3c8fdcb - [Analysis][ARM] Remove unused getNumBytesToPadGlobalArray (NFC) (#230912)
via llvm-commits
- [llvm] 762c28a - [llubi] Fix wrong assert when writing `poison` at an unaligned bit offset (#230819)
via llvm-commits
- [llvm] [llubi] Add support for `bitinsert` and `bitextract` (PR #230827)
Pedro Lobo via llvm-commits
- [llvm] d7bb842 - [llubi] Use correct tag bitwidth to recover provenances (#230852)
via llvm-commits
- [llvm] a408737 - [llubi] Reset retval for noop inline asm (#230849)
via llvm-commits
- [llvm] 93a5597 - [Verifier][IR] Make sure bitcast doesn't change type size (#230914)
via llvm-commits
- [clang] [llvm] [clang][CodeGenCoroutine] Fix throwing in unhandled_exception (PR #230933)
Weibo He via llvm-commits
- [llvm] [llvm][Coro] Lower coro.frame in CoroEarly instead of CoroSplit (PR #137976)
Weibo He via llvm-commits
- [clang] [llvm] [CoroEarly] Infer presplitcoroutine attribute for Switch-Resumed ABI (PR #169866)
Weibo He via llvm-commits
- [llvm] e5381bf - CodeGen: Merge TargetLoweringObjectFile::getModuleMetadata into Initialize (#226837)
via llvm-commits
- [llvm] [SPIR-V] Fix crash on byte buffer vector access at dynamic offset (PR #230934)
Arseniy Obolenskiy via llvm-commits
- [llvm] [AArch64] Prevent debug info from changing stack tag slot selection (PR #230935)
Yongqiang Tian via llvm-commits
- [llvm] [openmp] [OMPT][Offload][OpenMP] Fixes for OMPT data used by libomptarget (reapply #156020) (PR #175165)
via llvm-commits
- [llvm] 37dcd2c - [VPlan] Remove unused recipe constructor overloads (NFC) (#230857)
via llvm-commits
- [llvm] [llubi] Don't ignore address taken by assume-like intrinsics (PR #230938)
Yingwei Zheng via llvm-commits
- [llvm] [WinEHPrepare] Preserve PHI values introduced by spill stores (PR #230942)
Colen Garoutte-Carson via llvm-commits
- [llvm] 376a07d - [ValueTracking] Handle trunc to i1 condition in isKnownNonZeroFromAssume (#230919)
via llvm-commits
- [llvm] b7ae853 - [ValueTracking] Handle trunc to i1 condition in isKnownNonNullFromDominatingCondition (#230925)
via llvm-commits
- [llvm] eb61ef6 - [SCEV] Remove unused code from ScalarEvolutionExpressions.h (NFC) (#230859)
via llvm-commits
- [llvm] [DAGCombiner] Replace the node before the old load in reduceLoadWidth (PR #230947)
Aochang Liu via llvm-commits
- [llvm] 694d1a6 - [LAA] Compute MaxBTC * MaxStride in a wide enough type. (#230860)
via llvm-commits
- [llvm] [RISCV] Return the chain for strict FP_TO_FP16 on RV64 (PR #230948)
Aochang Liu via llvm-commits
- [llvm] 62487d7 - [InstCombine] fold ((X + C) & ~C) - X -> X & C where C is a power of 2 (#230796)
via llvm-commits
- [llvm] f4601bd - [llubi] Don't ignore address taken by assume-like intrinsics (#230938)
via llvm-commits
- [llvm] [X86][StackClash] Fix guard page skip with --fstack-clash-protection (PR #230951)
Shonie Caplan via llvm-commits
- [llvm] [LoopIdiom] Fix popcount trip count truncation and signed overflow (PR #230952)
Yunus Ozen via llvm-commits
- [llvm] [X86] Allow bf16 masked load/store with AVX512BW (PR #230953)
Divyansh Yadav via llvm-commits
- [clang] [llvm] [Clang][CUDA] Add support for __managed__ variables in non-RDC and default RDC mode (PR #149716)
via llvm-commits
- [llvm] [SelectionDAG] Optimize unaligned load stores to realign after offset (PR #145309)
via llvm-commits
- [llvm] 9a41343 - [MISched] Unify SUnit formatting among users (#229756)
via llvm-commits
- [llvm] [InstCombine] Don't fold `store undef` at non-byte-sized widths (PR #230955)
Mitch Briles via llvm-commits
- [llvm] [SelectionDAG] Fix fcmp fold for new min/max semantics (PR #223655)
Mitch Briles via llvm-commits
- [llvm] [SLP][NFC]Add a test with incorrect vectorization, NFC (PR #230958)
Alexey Bataev via llvm-commits
- [llvm] 092914e - [SLP][NFC]Add a test with incorrect vectorization, NFC
via llvm-commits
- [llvm] [SLP]Fix miscompile of absorbing lanes in flattened chains (PR #230959)
Alexey Bataev via llvm-commits
- [llvm] ea5ee40 - [SLP]Fix miscompile of absorbing lanes in flattened chains
via llvm-commits
- [llvm] cd9c8d0 - [Alignment] Add a static function to construct an `Align` from power-of-2 values (#230920)
via llvm-commits
- [llvm] [RISCV][P-ext] Move scalar mulh intrinsic tests to rv32p.ll and rv64p.ll. NFC (PR #230716)
Hongyu Chen via llvm-commits
- [clang] [flang] [libcxx] [llvm] [docs] Update obsolete Phabricator references for GitHub workflows (PR #229542)
Mehdi Amini via llvm-commits
- [llvm] f365350 - [CodeGenPrepare] Handle negative GEP offsets in splitLargeGEPOffsets (#227627)
via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Alexey Bataev via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Alexey Bataev via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Alexey Bataev via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Alexey Bataev via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Ryan Buchner via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
via llvm-commits
- [llvm] [SLP] Allow UDiv X, C <--> LShr X, log2(C) transformations in BinOpSameOpcodeHelper (PR #181731)
Ryan Buchner via llvm-commits
- [libc] [llvm] [libc] Make totalorder functions use the emulated float128 type (PR #221085)
Muhammad Haris Awan via llvm-commits
- [llvm] [unittests] Fix shared build. NFC (PR #230963)
via llvm-commits
- [llvm] [AArch64MIPeepholeOpt] Coalesce sibling base-address materializations (PR #223682)
via llvm-commits
- [llvm] [SelectionDAG] Preserve demanded lanes in sign-bit queries (PR #230093)
Simon Pilgrim via llvm-commits
- [llvm] [BPF] Treat class records as structs in BTF and CO-RE (PR #226723)
Maxim Skokov via llvm-commits
- [llvm] [MLGO][RegAlloc] Update test expectations after #228618 (PR #230973)
Mircea Trofin via llvm-commits
- [llvm] 4a59f39 - [MLGO][RegAlloc] Update test expectations after #228618 (#230973)
via llvm-commits
- [llvm] [PowerPC] Honor zeroext on widened constant returns in FastISel (PR #230977)
Manjunath S Matti. via llvm-commits
- [llvm] 7f6afc5 - [unittests] Fix shared build. NFC (#230963)
via llvm-commits
- [llvm] e170537 - [docs] Update obsolete Phabricator references for GitHub workflows (#229542)
via llvm-commits
- [llvm] [NFC][AMDGPU] Add a test showing unreachable call cycles in module splitting (PR #230983)
Shilei Tian via llvm-commits
- [llvm] 6c27fc9 - [InstCombine] Fix InstCombine pass to sink llvm.assume calls (#229658)
via llvm-commits
- [llvm] AMDGPU: Move scheduling model defs to AMDGPUSchedModels.td (PR #230995)
Matt Arsenault via llvm-commits
- [llvm] [InstCombine] Reuse existing xor for xor of disjoint or with constant (PR #230999)
via llvm-commits
- [libcxx] [llvm] [libc++] Updates Unicode tables to Unicode 18 (PR #199111)
Charlie Lin via llvm-commits
- [llvm] [CodeView] Add `CoroutineKind` to `S_FRAMEPROC` flags (PR #223018)
Alexandre Ganea via llvm-commits
- [llvm] X86/GlobalISel: Mark unused FP compare flag defs dead (PR #231003)
Matt Arsenault via llvm-commits
- [llvm] AMDGPU/GlobalISel: Mark scc dead on call frame and pc-relative pseudos (PR #231008)
Matt Arsenault via llvm-commits
- [llvm] [CodeView] Write 128bit ints as LF_(U)OCTWORD (PR #212241)
Alexandre Ganea via llvm-commits
- [llvm] 52bdac3 - [InstCombine] Remove redundant CRC32 intrinsic declarations from tests (#230926)
via llvm-commits
- [llvm] 201f3aa - [NFC][AMDGPU] Precommit test for byte lane extension after vector split (#228423)
via llvm-commits
- [llvm] [AMDGPU] Extract byte lanes of a split vector from the 32-bit source (PR #228424)
Dmitry Sidorov via llvm-commits
- [libc] [llvm] [libc] Make exponent decomposition f128 functions use emulated Float128 (PR #229855)
Mohammed Sohail via llvm-commits
- [llvm] Rename TableGen MIPS Registers to R<#> (PR #230696)
Anton Rapetov via llvm-commits
- [llvm] a8482df - [KnownFPClass] Fix `KnownFPClass::log` deductions for ppcf128 (#223296)
via llvm-commits
- [llvm] ecc1524 - AMDGPU: Move scheduling model defs to AMDGPUSchedModels.td (#230995)
via llvm-commits
- [llvm] AMDGPU: Move GCN subtarget features to GCNFeatures.td (PR #230996)
Matt Arsenault via llvm-commits
- [llvm] 1ca54c1 - [KnownFPClass] Apply denormal modes at the fmul/fdiv boundaries (#229638)
via llvm-commits
- [llvm] c208df9 - X86/GlobalISel: Mark unused FP compare flag defs dead (#231003)
via llvm-commits
- [compiler-rt] Repro_release_origin.c (PR #231012)
Vitaly Buka via llvm-commits
- [compiler-rt] [test] Remove empty --correlate from instrprof-binary-correlate-debuginfod.c (PR #231013)
Fangrui Song via llvm-commits
- [compiler-rt] 556e759 - [test] Remove empty --correlate from instrprof-binary-correlate-debuginfod.c (#231013)
via llvm-commits
- [llvm] 0b68182 - AMDGPU/GlobalISel: Mark scc dead on call frame and pc-relative pseudos (#231008)
via llvm-commits
- [llvm] 3ea0173 - [InstCombine] Don't fold `store undef` at non-byte-sized widths (#230955)
via llvm-commits
- [llvm] 18e8cfa - [LangRef] Fix typos in float min/max intrinsic docs (#224214)
via llvm-commits
- [llvm] [LangRef] Fix typos in float min/max intrinsic docs (PR #224214)
Yingwei Zheng via llvm-commits
- [llvm] [Hexagon] Ignore debug uses in constant extender optimization (PR #231014)
via llvm-commits
- [llvm] 99a9b94 - AMDGPU: Move GCN subtarget features to GCNFeatures.td (#230996)
via llvm-commits
- [llvm] caf743f - AMDGPU: Move R600 subtarget features to R600Features.td (#230997)
via llvm-commits
- [llvm] 738d73e - [AMDGPU] Extract byte lanes of a split vector from the 32-bit source (#228424)
via llvm-commits
- [llvm] TableGen: Split subtarget and processor classes out of Target.td (PR #231016)
Matt Arsenault via llvm-commits
- [llvm] test (PR #231020)
Namdar MohammadMehdi via llvm-commits
- [llvm] [libc][Bazel] Add missing float smoke math tests (PR #227206)
via llvm-commits
- [llvm] 9c67f5a - [RISCV][P-ext] Move scalar mulh intrinsic tests to rv32p.ll and rv64p.ll. NFC (#230716)
via llvm-commits
- [llvm] RISCV/GlobalISel: Mark unused implicit-def $x1 dead in call lowering (PR #231023)
Matt Arsenault via llvm-commits
- [llvm] a160747 - [Option] Add Condition for build-conditional library options (#230898)
via llvm-commits
- [llvm] [Verifier] Add type constraint for fpto[su]i.sat (PR #231024)
Yingwei Zheng via llvm-commits
- [llvm] [Vectorize] Declare command line options in TableGen (PR #231025)
Fangrui Song via llvm-commits
- [llvm] d429981 - TableGen: Split subtarget and processor classes out of Target.td (#231016)
via llvm-commits
- [llvm] AMDGPU: Use minimal tablegen inputs for TargetParser (PR #231017)
Matt Arsenault via llvm-commits
- [llvm] 5e68ef9 - Attributor: Use NullPointerIsDefined for null nocapture check (#229333)
via llvm-commits
- [llvm] RegAllocGreedy: Avoid unused block frequency lookups in hint recoloring (PR #231027)
Matt Arsenault via llvm-commits
- [llvm] [llubi] Use the correct poison representation for byte values (PR #231028)
Yingwei Zheng via llvm-commits
- [llvm] [AMDGPU] Encode op_sel_hi as zero for the V_PK 64-bit instructions (PR #230662)
Shilei Tian via llvm-commits
- [llvm] 5764334 - RISCV/GlobalISel: Mark unused implicit-def $x1 dead in call lowering (#231023)
via llvm-commits
- [llvm] [LoopBoundSplit] Preserve the original exit predicate for post-loop entry (PR #231029)
Yunus Ozen via llvm-commits
- [llvm] AMDGPU: Use TargetSubtarget.td in TargetDef tablegen tests (PR #231018)
Matt Arsenault via llvm-commits
- [llvm] 42c2dfb - [SelectionDAG] Verify SDTCisSameNumEltsAs in SDNodeInfo::verifyNode (#230616)
via llvm-commits
- [llvm] 36263cf - [RISCV][P-ext] Change RISCVISD::SATI to use [1,bitwidth] for the immediate. (#230377)
via llvm-commits
- [llvm] 007418c - RegAllocGreedy: Avoid unused block frequency lookups in hint recoloring (#231027)
via llvm-commits
- [llvm] 43e3e43 - [llubi] Use the correct poison representation for byte values (#231028)
via llvm-commits
- [llvm] [Passes] Run MergeICmps/ExpandMemCmps in LTO post-link pipeline (PR #230562)
Nikita Popov via llvm-commits
- [llvm] [GlobalISel] Call DSO-local nonlazybind callees directly (PR #231033)
via llvm-commits
- [llvm] [gn] make write_file.py replace literal \n with newlines (PR #231034)
Nico Weber via llvm-commits
- [llvm] [gn] port c90b6414e607 (mlgo EmitC models) (PR #231035)
Nico Weber via llvm-commits
- [llvm] 8a0968e - [gn] make write_file.py replace literal \n with newlines (#231034)
via llvm-commits
- [llvm] 6719b9e - [gn] port c90b6414e607 (mlgo EmitC models) (#231035)
via llvm-commits
- [clang] [llvm] [AArch64] Define apple-m5/a19 CPUs. (PR #171187)
Julien Poumailloux via llvm-commits
- [llvm] [GlobalISel] Improve bf16 converts with fast-math flags. (PR #200741)
David Green via llvm-commits
- [clang] [lld] [llvm] [WebAssembly] Emit constant initializers for addrspace(1) globals (PR #212007)
Max Desiatov via llvm-commits
- [llvm] [X86][GlobalISel] Select G_INSERT_VECTOR_ELT using SDAG patterns (PR #230317)
Evgenii Kudriashov via llvm-commits
- [llvm] [VPlan] Use getFastMathFlagsOrNone in VPIRFlags::applyFlags (NFC) (PR #231036)
Florian Hahn via llvm-commits
- [llvm] [LangRef] Clarify poison input elements give poison in llvm.mask.beforefirst (PR #230551)
Florian Hahn via llvm-commits
- [llvm] [X86][GlobalISel] Enable SDAG patterns for G_INSERT_VECTOR_ELT (PR #230317)
Evgenii Kudriashov via llvm-commits
- [llvm] [VPlan] Use erase_if in VPIRMetadata::intersect (NFC) (PR #231038)
Florian Hahn via llvm-commits
- [llvm] [LAA] Include the load types of written pointers in runtime check bounds. (PR #231039)
Florian Hahn via llvm-commits
- [llvm] [VPlan] Build packs from lane values in replicateByVF (NFC). (PR #231040)
Florian Hahn via llvm-commits
- [llvm] 5be1e6d - AMDGPU: Use TargetSubtarget.td in TargetDef tablegen tests (#231018)
via llvm-commits
- [llvm] 6239a74 - [VPlan] Use getFastMathFlagsOrNone in VPIRFlags::applyFlags (NFC) (#231036)
via llvm-commits
- [libc] [llvm] [libc][math] Optimize exp2 and refactor shared double-precision exp utilities. (PR #231041)
via llvm-commits
- [llvm] [SLP][NFC]Move poor throughput insts logic into a helper function, NFC (PR #231042)
Alexey Bataev via llvm-commits
- [llvm] af5dd51 - [SLP][NFC]Move poor throughput insts logic into a helper function, NFC
via llvm-commits
- [llvm] [orc-rt] Add Session instance symbols, SPS CI to look them up (PR #231043)
Lang Hames via llvm-commits
- [llvm] [Mips][GlobalISel] Ignore debug uses in register bank inference (PR #229948)
Jiaxun Yang via llvm-commits
- [llvm] [SDAG] Improve bf16 converts with fast-math flags. (PR #231044)
David Green via llvm-commits
- [llvm] 40ab028 - [VPlan] Build packs from lane values in replicateByVF (NFC). (#231040)
via llvm-commits
- [llvm] [LoopPeel] Do not set DesiredPeelCountLast in compare-elimination path (PR #231045)
via llvm-commits
- [llvm] 3853d20 - [VPlan] Use erase_if in VPIRMetadata::intersect (NFC) (#231038)
via llvm-commits
- [llvm] [MIR] Support +inf and -inf literals (PR #221127)
Taufiq Rahman via llvm-commits
- [compiler-rt] ce2a498 - [compiler-rt] Restrict __gnu_h2f_ieee and __gnu_f2h_ieee to arm32 EABI (#220170)
via llvm-commits
- [compiler-rt] [compiler-rt] Restrict __gnu_h2f_ieee and __gnu_f2h_ieee to arm32 EABI (PR #220170)
Saleem Abdulrasool via llvm-commits
- [llvm] [AArch64] Fold constants into their users after tail duplication (PR #231048)
Andrew Gaul via llvm-commits
- [llvm] [AMDGPU] Fix uniform fcopysign pattern (PR #218644)
via llvm-commits
- [llvm] [AMDGPU][X86][DAG] Avoid duplicate BinOp result from narrowing insert-extract sub-vector (PR #201056)
via llvm-commits
- [llvm] [KnownFPClass] Rule out non-zero finite results for fmul and fdiv (PR #231052)
Eric Ross via llvm-commits
- [llvm] Update [Github] Update GHA Dependencies (PR #231054)
Mend Renovate via llvm-commits
- [llvm] [LoongArch] Add a bitcast to Ext operand of VILVL in lowerVECTOR_SHUFFLEAsZeroOrAnyExtend. (PR #230607)
via llvm-commits
- [llvm] [Xtensa] Fix Cannot select when DAGCombine narrows a literal-pool load (PR #231056)
via llvm-commits
- [clang] [lld] [llvm] [RISC-V][RVY] Fix round-trip and order of arch strings with 'y' (PR #227865)
Alexander Richardson via llvm-commits
- [llvm] [MC] Add baseline test for MCContext::getSubtargetCopy HwMode loss (PR #227514)
Alexander Richardson via llvm-commits
- [lld] [lld][COFF] Mirror ELF LTO sample profile collection in reproducers (PR #225026)
Haohai Wen via llvm-commits
- [llvm] 9b9e3f0 - [test][llvm-ar] Restore file permissions in error-opening-permission.test (#230361)
via llvm-commits
- [llvm] [DAG] Fold fp_to_int(ftrunc(x)) -> fp_to_int(x) (PR #231060)
Adam Scott via llvm-commits
- [clang] [llvm] [LLVMABI][AARCH64] Implement general aggregate classification (PR #230631)
Madhur Amilkanthwar via llvm-commits
- [llvm] d1e15b6 - [BranchFolding] Ignore debug instructions in singleton cond branch check (#228715)
via llvm-commits
- [llvm] [mlir] [mlir][Arith] Move ceil/floor div expansion to ArithUtils. NFC (PR #231064)
Fangrui Song via llvm-commits
- [compiler-rt] [NFC][msan][test] Pin release_origin.c to a single CPU on Linux (PR #231058)
Vitaly Buka via llvm-commits
- [compiler-rt] [msan][test] Pin release_origin.c to a single CPU on Linux (PR #231058)
Vitaly Buka via llvm-commits
- [compiler-rt] [test][msan] Pin release_origin.c to a single CPU on Linux (PR #231058)
Vitaly Buka via llvm-commits
- [llvm] [AArch64] [DAG] combine vector_interleave to zip/zip2 (PR #231066)
Chen Zheng via llvm-commits
- [llvm] [MC][DWARF] Compare effective paths for root file reuse (PR #231067)
Ben Silver via llvm-commits
- [llvm] [CodeGen] Return an IR type from getCmpLibcallReturnType (NFC) (PR #227192)
Kiroo via llvm-commits
- [llvm] [NVPTX] Add atom.add.noftz support, avoid CAS loops when noftz is available (PR #223569)
Christian Sigg via llvm-commits
- [llvm] [CodeLayout] Avoid repeated full scans in ExtTSP merging (PR #214016)
Satoshi Kawasaki via llvm-commits
- [llvm] [LV] Convert scatter w/uniform addr and mask being header mask to scalar store. (PR #172799)
Elvis Wang via llvm-commits
- [llvm] 057f9bf - [AMDLIBM] Map new vector calls in VecFuncs (#224614)
via llvm-commits
- [llvm] [ObjectYAML][NFC] Hoist BBAddrMap obj2yaml decoder into a shared helper (PR #231071)
Haohai Wen via llvm-commits
- [llvm] [ARM][FastISel]Fix FastISel fail for softfp when set target to armebv… (PR #97422)
via llvm-commits
- [llvm] [InstCombine] Keep an alloca's metadata when folding its element count (PR #231073)
Pan Tao via llvm-commits
- [llvm] [IR] Remove unused debug intrinsic methods (NFC) (PR #231075)
Kazu Hirata via llvm-commits
- [llvm] [Attributor] Use sorted offsets for constant vector stores (PR #231079)
Jiahe Wang via llvm-commits
- [llvm] TableGen: Use DenseMap for the IntInit pool (PR #231082)
Matt Arsenault via llvm-commits
- [llvm] [X86] Don't double count the interrupt stack alignment push in StackSize (PR #224029)
Colin Finck via llvm-commits
Last message date:
Sun Oct 11 23:57:27 PDT 2026
Archived on: Sun Oct 11 23:57:35 PDT 2026
This archive was generated by
Pipermail 0.09 (Mailman edition).