Commits · e6012c8e0fbdf422325e66c96537de03d258deb7 · Roger Ferrer / llvm-epi · GitLab

Apr 07, 2022
- AMDGPU: Handle private atomics · e6012c8e
  Matt Arsenault authored Apr 05, 2022
  e6012c8e
- AtomicExpand: Add NotAtomic lowering strategy · 7f14a1d4
  Matt Arsenault authored Apr 05, 2022
  7f14a1d4
- AtomicExpand: Change return type for shouldExpandAtomicStoreInIR · c4ea925f
  Matt Arsenault authored Apr 05, 2022
  c4ea925f
- [RISCV] Supplement patterns for vnsrl.wx/vnsra.wx when splat shift is sext or zext · 1b547799
  Lian Wang authored Mar 31, 2022
  1b547799
- [AMDGPU] Check SI LDS offset bug in the allowsMisalignedMemoryAccesses · a41a676e
  Stanislav Mekhanoshin authored Apr 06, 2022
  a41a676e
- Transforms: Split LowerAtomics into separate Utils and pass · 39f15686
  Matt Arsenault authored Apr 05, 2022
  39f15686
- [MIPS] Initial support for MIPS-I load delay slots · 303c1801
  Simon Dardis authored Apr 07, 2022
  303c1801
- [MSSA] Print memory phis when inspecting walker. · 50d41f3e
  Alina Sbirlea authored Apr 06, 2022
  50d41f3e
- Revert f7381a79 · 08075a7e
  Alina Sbirlea authored Mar 29, 2022
  08075a7e
- [mips] Remove stale comment (NFC) · 8e1d9f00
  Simon Dardis authored Apr 06, 2022
  8e1d9f00
- [AArch64][AMDGPU][WebAssembly] Use static_cast instead of a reinterpret_cast... · 1235aaef
  Craig Topper authored Apr 06, 2022
  1235aaef
Apr 06, 2022
- [demangler] Node precision dumper · 4a4d0985
  Nathan Sidwell authored Mar 30, 2022
  4a4d0985
- [demangler][NFC] Rename SwapAndRestore to ScopedOverride · 51f6caf2
  Nathan Sidwell authored Feb 28, 2022
  51f6caf2
- [Support] [BLAKE3] Fix building for Cygwin · 9edee89b
  Martin Storsjö authored Apr 06, 2022
  9edee89b
- [AArch64] Fix the upper limit for folded address offsets for COFF · 8d7a17b7
  Martin Storsjö authored Apr 06, 2022
  8d7a17b7
- [LoopInterchange] Try to achieve the most optimal access pattern after interchange · eac34875
  Congzhe Cao authored Apr 06, 2022
  eac34875
- DebugInfo: Make the simplified template names prefix more unique · 6b306233
  David Blaikie authored Apr 06, 2022
  
  6b306233
- [LegalizeTypes][VP] Use LoVT/HiVT when splitting VP operations in SplitVecRes_UnaryOp. · bdb1ab98
  Craig Topper authored Apr 06, 2022
  bdb1ab98
- Revert "[CodeGen] Place SDNode debug ID declaration under appropriate #if" · 62a983eb
  Daniil Kovalev authored Apr 06, 2022
  62a983eb
- [LegalizeTypes] Move SplitVecRes_VECTOR_REVERSE/VECTOR_SPLICE near other SplitVecRes methods. NFC · 8fc19185
  Craig Topper authored Apr 06, 2022
  8fc19185
- [AArch64] Format the PerfectShuffle table. NFC · 560856d5
  David Green authored Apr 06, 2022
  
  560856d5
- [LegalizeDAG] Use SelectionDAG::getBoolConstant to simplify some code. NFC · 1ad36487
  Craig Topper authored Apr 06, 2022
  
  1ad36487
- [RISCV] Add lowering for vp.sext and vp.zext. · e13a44b4
  Craig Topper authored Apr 06, 2022
  e13a44b4
- [DAGCombiner] Replace call getSExtOrTrunc with a truncate. NFC · 5b5f5942
  Craig Topper authored Apr 06, 2022
  5b5f5942
- [DAGCombine] insert_subvector undef, (splat X), N2 -> splat X · 7d3af9ef
  Paul Walker authored Feb 22, 2022
  7d3af9ef
- [RISCV][VP] Add basic RVV codegen for vp.icmp · 6be5e875
  Fraser Cormack authored Mar 30, 2022
  6be5e875
- [X86] `lowerBuildVectorAsBroadcast()`: with AVX512VL, allow i64->XMM broadcasts from constant pool · 9be6e7b0
  Roman Lebedev authored Apr 06, 2022
  9be6e7b0
- [NFC] Remove redundant IndexType canonicalisation from DAGTypeLegalizer::PromoteIntOp_MSCATTER · 1c307b97
  Paul Walker authored Apr 06, 2022
  1c307b97
- [AArch64] Fold lsr+bfi in tryBitfieldInsertOpFromOr · 87f0d553
  chenglin.bi authored Apr 06, 2022
  87f0d553
- [SimplifyLibCalls] Use KnownBits helper APIs (NFC) · 1dc1d5a0
  Nikita Popov authored Apr 06, 2022
  1dc1d5a0
- MemoryBuiltins: getAllocAlignment is now useful for non-allocator funcs · 33b1f419
  Augie Fackler authored Mar 14, 2022
  33b1f419
- [AMDGPU] Fix unused variable warning after D117484 · 538c7717
  Jay Foad authored Apr 06, 2022
  
  538c7717
- [X86] Fold MMX_MOVD64from64rr + store to MMX_MOVQ64mr instead of... · f4661b5a
  Shengchen Kan authored Apr 06, 2022
  f4661b5a
- [SVE][AArch64] Enable first active true vector combine for INTRINSIC_WO_CHAIN · 9a2d5cc1
  zhongyunde authored Apr 06, 2022
  9a2d5cc1
- [AArch64][InstCombine] Fold MLOAD and zero extensions into MLOAD · 19e52351
  zhongyunde authored Apr 06, 2022
  19e52351
- [TLI] `TargetLowering::SimplifyDemandedVectorElts()`: narrowing bitcast: fill... · 34ce9fd8
  Roman Lebedev authored Apr 06, 2022
  34ce9fd8
- [CodeGen] Place SDNode debug ID declaration under appropriate #if · 83a798d4
  Daniil Kovalev authored Apr 06, 2022
  83a798d4
- [DebugInfo][InstrRef] Avoid a crash from mixed variable location modes · fb6596f1
  Jeremy Morse authored Apr 06, 2022
  fb6596f1
- Add support for more archs in `Triple::getArchTypeForLLVMName` · 060ff663
  Antonio Frighetto authored Apr 06, 2022
  060ff663
- [DAG] Allow XOR(X,MIN_SIGNED_VALUE) to perform AddLike folds · 3369e474
  Simon Pilgrim authored Apr 06, 2022
  3369e474