Commits · 81d0f17055e9f0b8c1d41606fc5bfc1d9e4126ea · Roger Ferrer / llvm-epi · GitLab

Jan 11, 2017
- Revert r291645 "[DAGCombiner] Teach DAG combiner to fold (vselect (N0 xor... · 7af39837
  Craig Topper authored Jan 11, 2017
  7af39837
- [DAGCombiner] Teach DAG combiner to fold (vselect (N0 xor AllOnes), N1, N2) ->... · 577d2585
  Craig Topper authored Jan 11, 2017
  577d2585
- Re-commit r289955: [X86] Fold (setcc (cmp (atomic_load_add x, -C) C), COND) to... · 6573976f
  Hans Wennborg authored Jan 11, 2017
  6573976f
- [X86] Dont run combineSetCCAtomicArith() when the cmp has multiple uses · 12de6937
  Hans Wennborg authored Jan 11, 2017
  12de6937
- [TM] Restore default TargetOptions in TargetMachine::resetTargetOptions. · 7d81813d
  Justin Lebar authored Jan 10, 2017
  7d81813d
- CodeGen: Allow small copyable blocks to "break" the CFG. · df27aa8c
  Kyle Butt authored Jan 10, 2017
  df27aa8c
Jan 10, 2017
- Make the test accept different OpCode values since it doesn't really care about the value. · ee787a76
  Douglas Yung authored Jan 10, 2017
  ee787a76
- DAG: Avoid OOB when legalizing vector indexing · 0b382a7c
  Matt Arsenault authored Jan 10, 2017
  0b382a7c
- [X86][AVX512]Improving shuffle lowering by using AVX-512 EXPAND* instructions · bcd03e7f
  Michael Zuckerman authored Jan 10, 2017
  bcd03e7f
- AMD family 17h (znver1) enablement · d55b8312
  Craig Topper authored Jan 10, 2017
  d55b8312
Jan 09, 2017
- [X86][AVX512VL] Added AVX512VL to 128/256 bit vector shift tests · fa328947
  Simon Pilgrim authored Jan 09, 2017
  fa328947
- PeepholeOptimizer: Do not replace SubregToReg(bitcast like) · ba7d95d4
  Matthias Braun authored Jan 09, 2017
  ba7d95d4
- Revert r291092 because it introduces a crash. · 1559e886
  Michael Kuperstein authored Jan 09, 2017
  1559e886
- X86-specific path: Implemented the fusing of MUL+ADDSUB to FMADDSUB. · d497d360
  Vyacheslav Klochkov authored Jan 09, 2017
  d497d360
- [X86][AVX512] Enable v16i8/v32i8 vector shifts to use an extend+shift+truncate pattern. · 0f23b2ba
  Simon Pilgrim authored Jan 09, 2017
  0f23b2ba
- [X86][AVX512DQ] Enable v16i16 vector shifts to use an extend+shift+truncate pattern. · d990cd37
  Simon Pilgrim authored Jan 09, 2017
  d990cd37
- [X86][AVX512DQ] Added AVX512DQ to 128/256 bit vector shift tests · f8538572
  Simon Pilgrim authored Jan 09, 2017
  f8538572
- [AVX-512] Change another pattern that was using BLENDM to use masked moves. A... · 96ab6fd2
  Craig Topper authored Jan 09, 2017
  96ab6fd2
- [AVX-512] Add patterns to use a zero masked VPTERNLOG instruction for vselects... · 6393afce
  Craig Topper authored Jan 09, 2017
  6393afce
Jan 08, 2017
- [AVX-512] If avx512dq is available use vpmovm2d/vpmovm2q instead of vselect of... · f51ba1e3
  Craig Topper authored Jan 08, 2017
  f51ba1e3
- [X86] Add avx512bw and avx512dq command lines to the vector compare results test. · 0930a523
  Craig Topper authored Jan 08, 2017
  0930a523
- [x86] fix usage of stale operands when lowering select · bf51c8a9
  Sanjay Patel authored Jan 08, 2017
  bf51c8a9
Jan 07, 2017
- [AVX-512] Remove patterns from the other VBLENDM instructions. They are all... · a74e3088
  Craig Topper authored Jan 07, 2017
  a74e3088
- [X86] Regenerate a test to remove tab characters. · 2d02d492
  Craig Topper authored Jan 07, 2017
  2d02d492
- [AVX-512] Add masked forms of the alternate MOVDDUP patterns. · da84ff3e
  Craig Topper authored Jan 07, 2017
  da84ff3e
- [X86][AVX2] Regenerate arithmetic tests · 935beac1
  Simon Pilgrim authored Jan 07, 2017
  935beac1
- [X86][AVX512] Use lowerShuffleAsRepeatedMaskAndLanePermute for non-VBMI v64i8 shuffles (PR31470) · a1b8e2c7
  Simon Pilgrim authored Jan 07, 2017
  a1b8e2c7
Jan 06, 2017
- [X86][SSE] Standardized triples in vector shift tests · 08519d7b
  Simon Pilgrim authored Jan 06, 2017
  08519d7b
- [X86][AVX] Regenerate shuffle 128-bit tests. · 9122793b
  Simon Pilgrim authored Jan 06, 2017
  9122793b
- [X86][AVX] Regenerate tzcnt tests. · 10cc5d55
  Simon Pilgrim authored Jan 06, 2017
  10cc5d55
- [AVX-512] Add EXTRACT_SUBVECTOR support to combineBitcastForMaskedOp. · e86fb932
  Craig Topper authored Jan 06, 2017
  e86fb932
- [AVX-512] Add more masked vector extract test cases with and without a bitcast between the select. · 8cbac879
  Craig Topper authored Jan 06, 2017
  8cbac879
Jan 05, 2017
- CodeGen: Assert that liveness is up to date when reading block live-ins. · 11723322
  Matthias Braun authored Jan 05, 2017
  11723322
- [x86] add test to show bug in select lowering; NFC · 686527c1
  Sanjay Patel authored Jan 05, 2017
  686527c1
- [X86] Add test cases that cover pr31551. NFC. · b10f7de3
  Zvi Rackover authored Jan 05, 2017
  b10f7de3
- [X86] Optimize vector shifts with variable but uniform shift amounts · 4b7d724d
  Zvi Rackover authored Jan 05, 2017
  4b7d724d
- [GlobalISel] Add support for address-taken basic blocks · a983e7c4
  Kristof Beyls authored Jan 05, 2017
  a983e7c4
- AVX-512: Optimized pattern for truncate with unsigned saturation. · 143cbc42
  Elena Demikhovsky authored Jan 05, 2017
  143cbc42
- [AVX-512] Update vextract64x4 intrinsic upgrade test cases to use a legal... · eea52429
  Craig Topper authored Jan 05, 2017
  eea52429
Jan 04, 2017
- The patch fixes (base, index, offset) match. · c88697dc
  Evgeny Stupachenko authored Jan 04, 2017
  c88697dc