Commits · cce3d96bcc6585e5f134fc093ec50813659c5c5e · Roger Ferrer / llvm-epi · GitLab

Mar 29, 2020
- GlobalISel: Add matcher for G_SHL · cce3d96b
  Matt Arsenault authored Mar 29, 2020
  
  cce3d96b
- AMDGPU/GlobalISel: Remove redundant virtual · d15723ef
  Matt Arsenault authored Mar 28, 2020
  
  d15723ef
- AMDGPU: Fix using wrong instruction for FP conversion · ab7a4106
  Matt Arsenault authored Mar 29, 2020
  ab7a4106
- AMDGPU: Add some additional tests for v_cvt_ubyte* formation · 0b68ca51
  Matt Arsenault authored Mar 29, 2020
  0b68ca51
- AMDGPU: Fix typo · 97bbe7ad
  Matt Arsenault authored Mar 29, 2020
  
  97bbe7ad
- [VectorCombine] skip debug intrinsics first for efficiency · fc3cc8a4
  Sanjay Patel authored Mar 29, 2020
  
  fc3cc8a4
- [InstCombine] make test independent of branch undef/UB; NFC · febcb24f
  Sanjay Patel authored Mar 29, 2020
  
  febcb24f
- [X86][AVX] Add tests for 512-bit shuffle patterns that could reduce to subvector extractions · 443dcc0e
  Simon Pilgrim authored Mar 29, 2020
  
  443dcc0e
- Remove unnecessary empty comments from test check lines. NFC. · b44f0704
  Simon Pilgrim authored Mar 29, 2020
  
  b44f0704
- [InstCombine] Simplify select of cmpxchg transform · 26fa3375
  Nikita Popov authored Mar 29, 2020
  26fa3375
- [OpenMP] set_bits iterator yields unsigned elements, no reference (NFC). · 99913ef3
  Florian Hahn authored Mar 29, 2020
  99913ef3
- [InstCombine] Fix worklist management in varargs transform · 28f67bd5
  Nikita Popov authored Mar 29, 2020
  28f67bd5
- [InstCombine] Erase original add when creating saddo · 6f07a9e8
  Nikita Popov authored Mar 29, 2020
  6f07a9e8
- [InstCombine] Use replaceOperand() in a few more places · 1e363023
  Nikita Popov authored Mar 29, 2020
  1e363023
- [X86][AVX] Combine 128-bit lane shuffles with a zeroable upper half to EXTRACT_SUBVECTOR (PR40720) · 7734e4b3
  Simon Pilgrim authored Mar 29, 2020
  7734e4b3
- [X86] Rename matchShuffleAsByteRotate to matchShuffleAsElementRotate. NFC. · da4c7db7
  Simon Pilgrim authored Mar 29, 2020
  da4c7db7
- [X86][AVX] Add X86ISD::VALIGN target shuffle decode support · 10439f9e
  Simon Pilgrim authored Mar 29, 2020
  10439f9e
- [VPlan] Use one VPWidenRecipe per original IR instruction. (NFC). · 49d00824
  Florian Hahn authored Mar 29, 2020
  49d00824
- [PostOrderIterator] Use SmallVector to store stack; NFC · 6ba63510
  Nikita Popov authored Mar 29, 2020
  6ba63510
- [X86] X86CallFrameOptimization - generalize slow push code path · a7115d51
  Simon Pilgrim authored Mar 29, 2020
  a7115d51
- [AlignmentFromAssumptions] Fix a SCEV assertion resulting from address space differences. · 4bf015c0
  Richard Diamond authored Mar 02, 2020
  4bf015c0
- [X86] Add cost model test cases for fmin/fmax reduction. · c0aa97b6
  Craig Topper authored Mar 28, 2020
  
  c0aa97b6
- [MC][PowerPC] Make .reloc support arbitrary relocation types · fc93787d
  Fangrui Song authored Mar 28, 2020
  fc93787d
Mar 28, 2020
- AMDGPU: Make use of default operands · 9564f467
  Matt Arsenault authored Sep 07, 2019
  
  9564f467
- Put back initializers that were dropped in 0ab5b5b8 · dd030036
  Benjamin Kramer authored Mar 28, 2020
  dd030036
- [MDBuilder] Don't use stable sort for sorting integers. · ba2e72c5
  Benjamin Kramer authored Mar 28, 2020
  
  ba2e72c5
- [InstCombine] Remove unreachable blocks before DCE · 2215dcf1
  Nikita Popov authored Mar 28, 2020
  2215dcf1
- [InstCombine] Merge two functions; NFC · 97cc1275
  Nikita Popov authored Mar 28, 2020
  97cc1275
- [ADT] Automatically forward llvm::sort to array_pod_sort if safe · d3b6e1f1
  Benjamin Kramer authored Mar 28, 2020
  d3b6e1f1
- [AMDGPU] Stabilize sort order · 2d24d74b
  Benjamin Kramer authored Mar 28, 2020
  2d24d74b
- [BPF] support 128bit int explicitly in layout spec · ced0d1f4
  Yonghong Song authored Mar 21, 2020
  ced0d1f4
- Upgrade some instances of std::sort to llvm::sort. NFC. · 4065e921
  Benjamin Kramer authored Mar 28, 2020
  
  4065e921
- [CodeGen] Fix sinking local values in lpads with phis · e5bf5037
  Reid Kleckner authored Mar 28, 2020
  e5bf5037
- [InstCombine] Use replaceOperand() API in GEP transforms · 30d71210
  Nikita Popov authored Mar 28, 2020
  30d71210
- [InstCombine] Reduce code duplication in GEP of PHI transform; NFC · b1f78bae
  Nikita Popov authored Mar 28, 2020
  b1f78bae
- After 09158252 , fix build when -DLLVM_ENABLE_THREADS=OFF · 3ab3f3c5
  Alexandre Ganea authored Mar 28, 2020
  3ab3f3c5
- [InstCombine] Fix worklist management in foldXorOfICmps() · 672e8bfb
  Nikita Popov authored Feb 21, 2020
  672e8bfb
- [InstCombine] Change limit-max-iterations test case; NFC · 337b671b
  Nikita Popov authored Mar 28, 2020
  337b671b
- [CorrelatedValuePropagation] Remove redundant if statement in processSelect() · 03bc311a
  Enna1 authored Mar 28, 2020
  03bc311a
- [AsmPrinter] Emit .weak directive for weak linkage on COFF for symbols without a comdat · e6112a56
  Martin Storsjö authored Mar 28, 2020
  e6112a56