Commits · ed78aaf093280af5686e0eda82a12ca24eb72baa · Roger Ferrer / llvm-epi · GitLab

Jun 22, 2017
- Revert "Enable vectorizer-maximize-bandwidth by default." · b512e915
  Diana Picus authored Jun 22, 2017
  b512e915
- [InstCombine] Add test cases to demonstrate that and->xnor and or->xnor... · 71e2c161
  Craig Topper authored Jun 22, 2017
  71e2c161
- Enable vectorizer-maximize-bandwidth by default. · 014db29b
  Dehao Chen authored Jun 21, 2017
  014db29b
Jun 21, 2017
- Add a "probe-stack" attribute · ed54b4a7
  whitequark authored Jun 21, 2017
  ed54b4a7
- Do not inline recursive direct calls in sample loader pass. · 50f2aa19
  Dehao Chen authored Jun 21, 2017
  50f2aa19
- [x86] set the datalayout to match the RUN line triple; NFC · deed5791
  Sanjay Patel authored Jun 21, 2017
  deed5791
- [InstCombine] Add range metadata to cttz/ctlz/ctpop intrinsic calls based on known bits · 5b173f2b
  Craig Topper authored Jun 21, 2017
  5b173f2b
- [InstCombine] Don't let folding (select (icmp eq (and X, C1), 0), Y, (or Y,... · ae86cc72
  Craig Topper authored Jun 21, 2017
  ae86cc72
- [Reassociate] Support xor reassociating for splat vectors · cbac691c
  Craig Topper authored Jun 21, 2017
  cbac691c
- [SCEV] Make MulOpsInlineThreshold lower to avoid excessive compilation time · eac01d4c
  Max Kazantsev authored Jun 21, 2017
  eac01d4c
Jun 20, 2017
- AMDGPU: Allow vectorization of packed types · 67cd347e
  Matt Arsenault authored Jun 20, 2017
  67cd347e
- [x86] enable CGP memcmp() expansion for 2/4/8 byte sizes · 0656629b
  Sanjay Patel authored Jun 20, 2017
  0656629b
- [InstCombine] fix code/test comments for r305792; NFC · 4ccbd58d
  Sanjay Patel authored Jun 20, 2017
  4ccbd58d
- [InstCombine] try to canonicalize xor-of-icmps to and-of-icmps · adca825d
  Sanjay Patel authored Jun 20, 2017
  adca825d
Jun 19, 2017
- [PATCH] [PGO] Fixed cast operation in emIntrinsicVisitor::instrumentOneMemIntrinsic. · f731bde0
  Ana Pazos authored Jun 19, 2017
  f731bde0
- Improve profile-guided heuristics to use estimated trip count. · 9083547a
  Taewook Oh authored Jun 19, 2017
  9083547a
- [InstCombine] Make sure AddReachableCodeToWorklist sets MadeIRChange · 475fcd9c
  Bjorn Pettersson authored Jun 19, 2017
  475fcd9c
- Revert r304824 "Fix PR23384 (part 3 of 3)" · ca69fc1c
  Hans Wennborg authored Jun 19, 2017
  ca69fc1c
- [InstCombine] Cleanup some duplicated one use checks · a7529b68
  Craig Topper authored Jun 19, 2017
  a7529b68
- [Reassociate] Support some reassociation of vector xors · ef85498e
  Craig Topper authored Jun 19, 2017
  ef85498e
- [TRE] Improve code motion in TRE, use AA to tell whether a load can be moved... · b412831d
  Xin Tong authored Jun 19, 2017
  b412831d
- [SCEV] Teach SCEVExpander to expand BinPow · 35b2a18e
  Max Kazantsev authored Jun 19, 2017
  35b2a18e
- NewGVN: Fix PR 33461, caused by slightly overzealous verification. · 36b08b20
  Daniel Berlin authored Jun 19, 2017
  36b08b20
Jun 18, 2017
- Add argmononly attribute to strlen and wcslen, i.e. they only read memory (string) passed to them. · 9d2a5b1c
  Xin Tong authored Jun 18, 2017
  9d2a5b1c
Jun 17, 2017
- [SROA] Add support for non-integral pointers · b70ddd89
  Sanjoy Das authored Jun 17, 2017
  b70ddd89
- [InstCombine] Make FPMathOperator working with ConstantExpression(s). · 64f94fe0
  Davide Italiano authored Jun 17, 2017
  64f94fe0
- Revert rL305578. There is still some buildbot failure to be fixed. · c7ba8763
  Wei Mi authored Jun 16, 2017
  c7ba8763
Jun 16, 2017
- [InstCombine] Set correct insertion point for selects generated while folding phis · 6bc14c65
  Anna Thomas authored Jun 16, 2017
  6bc14c65
- Change YAML traits for vector<string> to flow_vector. · 69e2b1d4
  Evgeniy Stepanov authored Jun 16, 2017
  69e2b1d4
- [GVN] Recommit the patch "Add phi-translate support in scalarpre". · a2493b6a
  Wei Mi authored Jun 16, 2017
  a2493b6a
- [InstCombine] Add test cases to show missed opportunities due to overly... · 31399b70
  Craig Topper authored Jun 16, 2017
  31399b70
- [Atomics] Rename and change prototype for atomic memcpy intrinsic · 3faabbbe
  Daniel Neilson authored Jun 16, 2017
  3faabbbe
- [InstCombine] Fold (!iszero(A & K1) & !iszero(A & K2)) -> (A & (K1 | K2)) ==... · da6ea0d3
  Craig Topper authored Jun 16, 2017
  da6ea0d3
- [cfi] CFI-ICall for ThinLTO. · 4d4ee93d
  Evgeniy Stepanov authored Jun 16, 2017
  4d4ee93d
Jun 15, 2017
- [InstCombine] Add test cases to demonstrate instcombine increasing instruction... · ca73ba01
  Craig Topper authored Jun 15, 2017
  ca73ba01
- [InstCombine] Pre-commit test cases for the transform proposed in D34244. · 51dc5327
  Craig Topper authored Jun 15, 2017
  51dc5327
- [InstCombine] Handle (iszero(A & K1) | iszero(A & K2)) -> (A & (K1 | K2)) !=... · 6eec9e21
  Craig Topper authored Jun 15, 2017
  6eec9e21
- [BasicAA] Add test case that goes with r305481. · 6ec5a630
  Craig Topper authored Jun 15, 2017
  6ec5a630
- [InstCombine] auto-generate complete checks; NFC · 47496695
  Sanjay Patel authored Jun 15, 2017
  47496695
- [InstCombine] Add a test case to show a case where don't handle a partially commuted IR. NFC · 6ea89f24
  Craig Topper authored Jun 15, 2017
  6ea89f24