Commits · 255343483d452ac76638a02d49214728887b8430 · Lorenzo Albano / LLVM bpEVL · GitLab

Feb 11, 2017
- [AVX-512] Add VPMINS/MINU/MAXS/MAXU instructions to load folding tables. · 25534348
  Craig Topper authored Feb 11, 2017
  25534348
- [X86] Improve alphabetizing of load folding tables. NFC · b2fa216d
  Craig Topper authored Feb 11, 2017
  b2fa216d
- [X86][SSE] Convert getTargetShuffleMaskIndices to use getTargetConstantBitsFromNode. · 0e6945e4
  Simon Pilgrim authored Feb 11, 2017
  0e6945e4
- docs: update docs for objc_storeStrong behaviour · 5b1f0edf
  Saleem Abdulrasool authored Feb 11, 2017
  5b1f0edf
- CodeGen: rename variables to adhere to naming convention · e60561c0
  Saleem Abdulrasool authored Feb 11, 2017
  e60561c0
- Sema: simplify conditional execution (NFC) · b893ed26
  Saleem Abdulrasool authored Feb 11, 2017
  b893ed26
- [X86] Merge repeated getScalarValueSizeInBits calls. NFCI. · d59fa0e3
  Simon Pilgrim authored Feb 11, 2017
  d59fa0e3
- NewGVN: Reverse sense of this test to make it clearer · 22a4a01f
  Daniel Berlin authored Feb 11, 2017
  22a4a01f
- NewGVN: Add missing initialization of NumFuncArgs lost due to bad merge. · 1529bb93
  Daniel Berlin authored Feb 11, 2017
  1529bb93
- NewGVN: Rank and order commutative operands consistently. · 1c08767f
  Daniel Berlin authored Feb 11, 2017
  1c08767f
- [X86][3DNow!] Add tests to ensure PFMAX/PFMIN are not commuted. · 86a95c1f
  Simon Pilgrim authored Feb 11, 2017
  86a95c1f
- [X86][3DNow!] Enable PFSUB<->PFSUBR commutation · 6411a0eb
  Simon Pilgrim authored Feb 11, 2017
  6411a0eb
- [X86][3DNow!] Enable commutation for PFADD/PFMUL/PFCMPEQ/PAVGUSB/PMULHRW · 4ead1d4a
  Simon Pilgrim authored Feb 11, 2017
  4ead1d4a
- [X86][3DNow!] Add tests showing missed commutation opportunities. · 6b4a5134
  Simon Pilgrim authored Feb 11, 2017
  6b4a5134
- NewGVN: Clean up how we handle the INITIAL class so that everything in · b79f5366
  Daniel Berlin authored Feb 11, 2017
  b79f5366
- Fix "left shift of negative value -1" introduced by r294805 · bcb6622c
  Vitaly Buka authored Feb 11, 2017
  bcb6622c
- This reverts commits r294826 and r294781 as they break linking on powerpc. · d8230247
  Vitaly Buka authored Feb 11, 2017
  d8230247
- [X86][XOP] Regenerate XOP commutation tests. · 8158816e
  Simon Pilgrim authored Feb 11, 2017
  8158816e
- [X86][SSE] Regenerate float comparison commutation tests. · 008ba63e
  Simon Pilgrim authored Feb 11, 2017
  008ba63e
- [X86] Regenerate CLMUL commutation tests. · 0d8632f0
  Simon Pilgrim authored Feb 11, 2017
  0d8632f0
- Make helpers static. NFC. · 357c9e1a
  Benjamin Kramer authored Feb 11, 2017
  357c9e1a
- Move symbols from the global namespace into (anonymous) namespaces. NFC. · efcf06f5
  Benjamin Kramer authored Feb 11, 2017
  efcf06f5
- Check reduction dependencies in case of the matrix multiplication optimization · b196055c
  Roman Gareev authored Feb 11, 2017
  b196055c
- [FIX] Fix the potential issue of containsOnlyMatMulDep. · de69293b
  Roman Gareev authored Feb 11, 2017
  de69293b
- [NFC] Fix the style issue of lib/Transform/ScheduleOptimizer.cpp. · 5ef7e210
  Roman Gareev authored Feb 11, 2017
  5ef7e210
- Remove a now unneeded __CloudABI__ check. · 252da3b3
  Ed Schouten authored Feb 11, 2017
  252da3b3
- Fix the build of thread.cpp on CloudABI. · de5669e4
  Ed Schouten authored Feb 11, 2017
  de5669e4
- [NFC] Fix style issues of lib/Transform/ScheduleOptimizer.cpp. · afcf026d
  Roman Gareev authored Feb 11, 2017
  afcf026d
- [AVX-512] Add VPINSRB/W/D/Q instructions to load folding tables. · 1f6153ba
  Craig Topper authored Feb 11, 2017
  1f6153ba
- [AVX-512] Fix apparent typo in instruction name VMOVSSDrr_REV->VMOVSDZrr_REV. · a9818aad
  Craig Topper authored Feb 11, 2017
  a9818aad
- Use the size of the widest type of the matrix multiplication operands · 3d4eae31
  Roman Gareev authored Feb 11, 2017
  3d4eae31
- [AVX-512] Add VPSADBW instructions to load folding tables. · 3afa777f
  Craig Topper authored Feb 11, 2017
  3afa777f
- Fix -Wsign-compare - this might not be quite right, but preserves behavior · 9730789a
  David Blaikie authored Feb 11, 2017
  9730789a
- The patch fixes r294821 · 5f3d9b6c
  Evgeny Stupachenko authored Feb 11, 2017
  5f3d9b6c
- [X86] Don't base domain decisions on VEXTRACTF128/VINSERTF128 if only AVX1 is available. · 464b8cb2
  Craig Topper authored Feb 11, 2017
  464b8cb2
- Fix memory leak by using unique_ptr · a67cf000
  David Blaikie authored Feb 11, 2017
  a67cf000
- Address Mehdi's post-commit review comments on r294795. · fa3175f2
  Peter Collingbourne authored Feb 11, 2017
  fa3175f2
- Fix PR23384 (under "-lsr-insns-cost" option) · fe6f548d
  Evgeny Stupachenko authored Feb 11, 2017
  fe6f548d
- Update XFAIL line after r294781. · a05bdf75
  Benjamin Kramer authored Feb 11, 2017
  a05bdf75
- [ARM] Make f16 interleaved accesses expensive. · 8425f453
  Ahmed Bougacha authored Feb 11, 2017
  8425f453