Commits · d992950078efd81971b8d84a862671b17af88bff · Roger Ferrer / llvm-epi · GitLab

Sep 27, 2021
- [InstCombine] add tests for opposing shifts separated by trunc; NFC · d9929500
  Sanjay Patel authored Sep 27, 2021
  
  d9929500
- [InstCombine] generalize fold for (trunc (X u>> C1)) u>> C · 21429cf4
  Sanjay Patel authored Sep 27, 2021
  21429cf4
Sep 26, 2021
- [InstCombine] move add after min/max intrinsic · 6063e6b4
  Sanjay Patel authored Sep 26, 2021
  6063e6b4
Sep 25, 2021
- [InstCombine] Ensure shifts are in range for (X << C1) / C2 -> X fold. · 5a14edd8
  Simon Pilgrim authored Sep 25, 2021
  5a14edd8
Sep 24, 2021
- [InstCombine] fold lshr(trunc(lshr X, C1)) C2 · a47c8e40
  Sanjay Patel authored Sep 24, 2021
  a47c8e40
- [InstCombine] add tests for lshr-trunc-lshr; NFC · 638a4147
  Sanjay Patel authored Sep 24, 2021
  
  638a4147
- Revert "[InstCombine] fold cast of right-shift if high bits are not demanded (2nd try)" · 3c550090
  Sanjay Patel authored Sep 24, 2021
  3c550090
Sep 23, 2021
- [InstCombine] fold cast of right-shift if high bits are not demanded (2nd try) · bb9333c3
  Sanjay Patel authored Sep 23, 2021
  bb9333c3
- [AArch64][SVE][InstCombine] Eliminate redundant chains of tuple get/set · 3b12282b
  Usman Nadeem authored Sep 13, 2021
  3b12282b
Sep 22, 2021
- [InstCombine] add one-use check to shift-shift transform · 1cd6b44f
  Sanjay Patel authored Sep 22, 2021
  1cd6b44f
- [InstCombine] add test for shift-shift with extra use; NFC · 55aa4e92
  Sanjay Patel authored Sep 22, 2021
  
  55aa4e92
- [ValueTracking] fix isOnlyUsedInZeroEqualityComparison with no users · a85d7a56
  Sanjay Patel authored Sep 22, 2021
  a85d7a56
- [Analysis] improve function matching for strlen libcall · c240169f
  Sanjay Patel authored Sep 22, 2021
  c240169f
- [InstCombine] Update InstCombine to use poison instead of undef for... · 98e96663
  hyeongyu kim authored Sep 23, 2021
  98e96663
- [InstCombine] Update InstCombine to use poison instead of undef for... · ec831144
  hyeongyu kim authored Sep 22, 2021
  ec831144
- [InstCombine] Update InstCombine to use poison instead of undef for... · e5aaf033
  hyeongyu kim authored Sep 22, 2021
  e5aaf033
- Revert "[InstCombine] fold cast of right-shift if high bits are not demanded" · c6013f71
  Sanjay Patel authored Sep 22, 2021
  c6013f71
- Don't fold (select C, (gep Ptr, Idx), Ptr) if C is vector but Idx is scalar · d0746f2e
  Yi Kong authored Sep 22, 2021
  d0746f2e
Sep 21, 2021
- [InstCombine] fold cast of right-shift if high bits are not demanded · 2f6b0731
  Sanjay Patel authored Sep 21, 2021
  2f6b0731
- Teach InstCombine to eliminate malloc-realloc-free triplets. · b5fbbdd2
  Owen Anderson authored Sep 17, 2021
  b5fbbdd2
- [InstCombine] powi(x, y) * powi(x, z) -> powi(x, y + z) · c0fdfc9a
  Dávid Bolvanský authored Sep 21, 2021
  c0fdfc9a
- [InstCombine] move/add tests for trunc-of-lshr; NFC · 08ef71ca
  Sanjay Patel authored Sep 21, 2021
  08ef71ca
- [InstCombine] Improve TryToSinkInstruction with multiple uses · 69921f6f
  Anna Thomas authored Sep 20, 2021
  69921f6f
- [InstCombine] add tests for mask-shift with trunc; NFC · af1c5312
  Sanjay Patel authored Sep 20, 2021
  
  af1c5312
- [InstCombine] foldConstantInsEltIntoShuffle - bail if we fail to find constant element (PR51824) · fc8f1e44
  Simon Pilgrim authored Sep 21, 2021
  fc8f1e44
- [InstCombine] Eliminate vector reverse if all inputs/outputs to an instruction are reverses · f417d9d8
  Usman Nadeem authored Sep 20, 2021
  f417d9d8
Sep 20, 2021
- [IR] Add helper to convert offset to GEP indices · dd022656
  Nikita Popov authored Sep 19, 2021
  dd022656
- [Analysis] Add support for vscale in computeKnownBitsFromOperator · f988f680
  David Sherwood authored Sep 16, 2021
  f988f680
Sep 19, 2021
- [InstCombine] add/adjust tests for min/max intrinsics; NFC · 9555d1ed
  Sanjay Patel authored Sep 19, 2021
  9555d1ed
Sep 18, 2021
- [Tests] Fix noalias metadata in one more test · abe21da6
  Nikita Popov authored Sep 18, 2021
  abe21da6
- [Tests] Fix incorrect noalias metadata · 80110aaf
  Nikita Popov authored Sep 16, 2021
  80110aaf
- Precommit tests for D109807 "[InstCombine] Narrow type of logical operation chains when possible" · d841c72e
  Usman Nadeem authored Sep 18, 2021
  d841c72e
- [AArch64][SVE][InstCombine] Fold redundant zip1/2(uzp1/2) operations · 757384ab
  Usman Nadeem authored Sep 12, 2021
  757384ab
Sep 17, 2021
- [InstCombine] add tests for min/max intrinsics with offset operand; NFC · 6da35036
  Sanjay Patel authored Sep 17, 2021
  
  6da35036
- [NFC] Precommit tests for D109954 · d01e0c8c
  Dávid Bolvanský authored Sep 17, 2021
  
  d01e0c8c
- [InstCombine] allow splat vectors for narrowing masked fold · 41ff7612
  Sanjay Patel authored Sep 17, 2021
  41ff7612
- [InstCombine] add vector tests for 'and' folds; NFC · 3a587ed2
  Sanjay Patel authored Sep 17, 2021
  
  3a587ed2
Sep 16, 2021
- [InstCombine] Added llvm.powi optimizations · a4a426c9
  Dávid Bolvanský authored Sep 16, 2021
  a4a426c9
- [NFC] Added tests for llvm.powi optimizations · c0afb009
  Dávid Bolvanský authored Sep 16, 2021
  
  c0afb009
- Revert "[InstCombine] Improve TryToSinkInstruction with multiple uses" · f9e4aebe
  Anna Thomas authored Sep 15, 2021
  f9e4aebe