Commits · ff30436dc5e54b85b5b942a3a84d0720f657b36f · Roger Ferrer / llvm-epi · GitLab

May 25, 2021
- [GlobalISel] Fix MachineIRBuilder not using the DstOp argument for G_SHUFFLE_VECTOR. · ff30436d
  Amara Emerson authored May 24, 2021
  
  ff30436d
- [RISCV] Optimize xor/or with immediate in the zbs extension · bf773170
  Ben Shi authored May 25, 2021
  bf773170
- [LoopDeletion] Break backedge if we can prove that the loop is exited on 1st iteration · 2531fd70
  Max Kazantsev authored May 25, 2021
  2531fd70
- AMDGPU/GlobalISel: Legalize G_[SU]DIVREM instructions · 90d78405
  Christudasan Devadasan authored Apr 12, 2021
  90d78405
- [JITLink] Enable creation and management of mutable block content. · 82ad2b6e
  Lang Hames authored May 24, 2021
  82ad2b6e
- Revert "[LoopUnrollAndJam] Change LoopUnrollAndJamPass to LoopNest pass" · e77d24f7
  maekawatoshiki authored May 25, 2021
  e77d24f7
- [NFC][OMP] Fix 'unused' warning · 676a789a
  Vitaly Buka authored May 24, 2021
  
  676a789a
- [SLP] Fix "gathering" of insertelement instructions · b2cd8950
  Anton Afanasyev authored May 18, 2021
  b2cd8950
May 24, 2021
- [RISCV] Add a vsetvli insert pass that can be extended to be aware of incoming... · b510e4cf
  Craig Topper authored May 24, 2021
  b510e4cf
- [WebAssembly] Add NullifyDebugValueLists pass · a64ebb86
  Heejin Ahn authored May 23, 2021
  a64ebb86
- [X86] Call insertDAGNode on trunc/zext created in tryShiftAmountMod. · 3c0735c6
  Craig Topper authored May 24, 2021
  3c0735c6
- [Remarks] Add analysis remarks for memset/memcpy/memmove lengths · 095e91c9
  Jon Roelofs authored May 24, 2021
  095e91c9
- [X86][Costmodel] getMaskedMemoryOpCost(): don't scalarize non-power-of-two... · c666208f
  Roman Lebedev authored May 24, 2021
  c666208f
- [RISCV] Optimize getVLENFactoredAmount function. · d70e9195
  luxufan authored May 24, 2021
  d70e9195
- [Remarks] Look through inttoptr/ptrtoint for -ftrivial-auto-var-init remarks. · 694068d0
  Jon Roelofs authored May 24, 2021
  694068d0
- CoroSplit: Replace ad-hoc implementation of reachability with API from CFG.h · 4cba0a4f
  Adrian Prantl authored May 24, 2021
  4cba0a4f
- [CostModel][X86] Add missing SSE41 v2iX sext/zext costs · dcaca720
  Simon Pilgrim authored May 24, 2021
  dcaca720
- [NVPTX] Fix lowering of frem for negative values · 505933a4
  thomasraoux authored May 24, 2021
  505933a4
- [ConstProp] propagate poison from vector reduction element(s) to result · a0e71f18
  Sanjay Patel authored May 24, 2021
  a0e71f18
- [VPlan] Add first VPlan version of sinkScalarOperands. · 65d3dd7c
  Florian Hahn authored May 24, 2021
  65d3dd7c
- [CostModel][X86] Improve accuracy of vector non-uniform shift costs on XOP/AVX2 targets · 1ad4f887
  Simon Pilgrim authored May 24, 2021
  1ad4f887
- [VPlan] Add mayReadOrWriteMemory & friends. · e9d97d7d
  Florian Hahn authored May 14, 2021
  e9d97d7d
- [AArch64][SVE] Add fixed length codegen for FP_ROUND/FP_EXTEND · e4051325
  Bradley Smith authored May 11, 2021
  e4051325
- [AArch64][SVE] Improve codegen for fixed length vector concat · 4bc14be2
  Bradley Smith authored May 14, 2021
  4bc14be2
- [ARM] Allow findLoopPreheader to return headers with multiple loop successors · 543406a6
  David Green authored May 24, 2021
  543406a6
- Recommit "[VectorCombine] Scalarize vector load/extract." · 4e8c28b6
  Florian Hahn authored May 24, 2021
  4e8c28b6
- [ARM] Ensure WLS preheader blocks have branches during memcpy lowering · 53c42f77
  David Green authored May 24, 2021
  53c42f77
- [ARM] Fix inline memcpy trip count sequence · 6cc78b92
  David Green authored May 24, 2021
  6cc78b92
- [RISCV] Prevent store combining from infinitely looping · 7a211ed1
  Fraser Cormack authored May 21, 2021
  7a211ed1
- [NFCI][LoopIdiom] 'left-shift until bittest': assert that BaseX is loop-invariant · 32bee427
  Roman Lebedev authored May 24, 2021
  32bee427
- [LoopIdiom] 'logical right shift until zero': the value must be loop-invariant · aa3dac95
  Roman Lebedev authored May 24, 2021
  aa3dac95
- Revert "[VectorCombine] Scalarize vector load/extract." · 94d54155
  Florian Hahn authored May 24, 2021
  94d54155
- [CostModel][X86] Improve accuracy of vXi64 MUL costs on AVX2/AVX512 targets · 243e5886
  Simon Pilgrim authored May 23, 2021
  243e5886
- [VectorCombine] Scalarize vector load/extract. · 86497785
  Florian Hahn authored May 24, 2021
  86497785
- [Attributor] Introduce a helper do deal with constant type mismatches · 6caea8a7
  Johannes Doerfert authored May 09, 2021
  6caea8a7
- [Attributor] Teach AAIsDead about undef values · 55e9c282
  Johannes Doerfert authored May 07, 2021
  55e9c282
- [Attributor] Deal with address spaces gracefully · 4878d734
  Johannes Doerfert authored May 06, 2021
  4878d734
- [Attributor] Be more careful to not disturb the CG outside the SCC · 1ba2929b
  Johannes Doerfert authored May 15, 2021
  1ba2929b
- [Attributor][FIX] Account for undef in the constant value lattice · e93ac1e2
  Johannes Doerfert authored May 07, 2021
  e93ac1e2
- [Attributor][FIX] Ensure we replace undef if we see the first "real" value · 5cdc29f7
  Johannes Doerfert authored May 06, 2021
  5cdc29f7