Commits · fd7300f717c18c861e77685efe6f16f12fb63ae7 · Roger Ferrer / llvm-epi · GitLab

Mar 18, 2020
- [NFC][PowerPC] Update test · fc2a5ef9
  Sam Parker authored Mar 18, 2020
  fc2a5ef9
Mar 12, 2020
- [LoopPeel] Turn incorrect assert into a check · 3dc6e53c
  Max Kazantsev authored Mar 12, 2020
  3dc6e53c
Mar 06, 2020
- [InstComine] Forego of one-use check in `(X - (X & Y)) --> (X & ~Y)` if Y is a constant · 1badf7c3
  Roman Lebedev authored Mar 06, 2020
  1badf7c3
Mar 02, 2020
- [Loop Peeling] Add possibility to enable peeling on loop nests. · 3dcaf296
  Arkady Shlykov authored Jan 17, 2020
  3dcaf296
Jan 23, 2020
- [LoopUnroll] Avoid UB when converting from WeakVH to `Value *` · b81a337b
  Justin Bogner authored Jan 23, 2020
  b81a337b
Jan 20, 2020
- [LoopUtils] Better accuracy for getLoopEstimatedTripCount. · 10357e1c
  Evgeniy Brevnov authored Jan 16, 2020
  10357e1c
Jan 16, 2020
- Revert "[Loop Peeling] Add possibility to enable peeling on loop nests." · c87982b4
  Arkady Shlykov authored Jan 16, 2020
  c87982b4
- [llvm] Make new pass manager's OptimizationLevel a class · 7acfda63
  Mircea Trofin authored Jan 16, 2020
  7acfda63
Jan 15, 2020
- [Loop Peeling] Add possibility to enable peeling on loop nests. · 3f3017e1
  Arkady Shlykov authored Jan 15, 2020
  3f3017e1
- [NFC] Adjust test cases numbering, test commit. · 019c8d9d
  Arkady Shlykov authored Jan 14, 2020
  019c8d9d
Jan 10, 2020
- Follow up of 67bf9a61 , minor fix in test case, removed duplicate option · 356685a1
  Sjoerd Meijer authored Jan 10, 2020
  
  356685a1
- [SVEV] Recognise hardware-loop intrinsic loop.decrement.reg · 67bf9a61
  Sjoerd Meijer authored Jan 10, 2020
  67bf9a61
Jan 09, 2020
- [ARM][MVE] Don't unroll intrinsic loops. · 15c7fa4d
  Sam Parker authored Jan 09, 2020
  15c7fa4d
Dec 25, 2019
- Migrate function attribute "no-frame-pointer-elim" to "frame-pointer"="all" as... · 502a77f1
  Fangrui Song authored Dec 24, 2019
  502a77f1
Dec 02, 2019
- [InstCombine] Revert rL341831: relax one-use check in foldICmpAddConstant() (PR44100) · 0f22e783
  Roman Lebedev authored Dec 02, 2019
  0f22e783
Nov 19, 2019
- [AMDGPU] Tune inlining parameters for AMDGPU target (part 2) · 6fd11b14
  dfukalov authored Nov 18, 2019
  6fd11b14
Nov 07, 2019
- [LoopPred] Enable new transformation by default · 8748be77
  Philip Reames authored Nov 06, 2019
  8748be77
Nov 06, 2019
- [LoopUnroll] countToEliminateCompares(): fix handling of [in]equality predicates (PR43840) · 4fe94d03
  Roman Lebedev authored Nov 06, 2019
  4fe94d03
- [NFC][LoopUnroll] Update test coverage for peeling w/ inequality predicates · 432a12c8
  Roman Lebedev authored Nov 06, 2019
  
  432a12c8
- [AMDGPU] Improve code size cost model (part 2) · 47a5c36b
  dfukalov authored Nov 05, 2019
  47a5c36b
Nov 05, 2019
- [LoopUnroll] peel-loop-conditions.ll: add some 'is even/odd' peeling tests · 12c4a71c
  Roman Lebedev authored Nov 05, 2019
  
  12c4a71c
Oct 30, 2019
- [NFC][LoopUnroll] Tests for peeling of first iteration (PR43840) · 0405b486
  Roman Lebedev authored Oct 30, 2019
  
  0405b486
Oct 29, 2019
- [LCSSA] Forget values we create LCSSA phis for · 596e4ab9
  Florian Hahn authored Oct 29, 2019
  596e4ab9
Sep 26, 2019
- [Unroll] Do NOT unroll a loop with small runtime upperbound · 1128fa09
  Zhaoshi Zheng authored Sep 26, 2019
  1128fa09
Sep 19, 2019
- [Unroll] Add an option to control complete unrolling · a4476885
  Serguei Katkov authored Sep 19, 2019
  a4476885
Sep 17, 2019
- [LoopUnroll] Use LoopSize+1 as threshold, to allow unrolling loops matching LoopSize. · 1bd58870
  Florian Hahn authored Sep 17, 2019
  1bd58870
Aug 26, 2019
- [LoopUnroll] Handle certain PHIs in full unrolling properly · d804bd17
  Bjorn Pettersson authored Aug 26, 2019
  d804bd17
Aug 22, 2019
- [Loop Peeling] Fix silly bug in metadata update. · 036e636a
  Serguei Katkov authored Aug 22, 2019
  036e636a
Aug 14, 2019
- [RLEV] Rewrite loop exit values for multiple exit loops w/o overall loop exit count · 6cca3ad4
  Philip Reames authored Aug 14, 2019
  6cca3ad4
Aug 11, 2019
- [MVE] Don't try to unroll vectorised MVE loops · 11c4602f
  David Green authored Aug 11, 2019
  11c4602f
Aug 02, 2019
- [Loop Peeling] Introduce an option for profile based peeling disabling. · de67affd
  Serguei Katkov authored Aug 02, 2019
  de67affd
- [Loop Peeling] Do not close further unroll/peel if profile based peeling was not used. · bbdcc821
  Serguei Katkov authored Aug 02, 2019
  bbdcc821
Jul 25, 2019
- [Loop Peeling] Fix idom detection algorithm. · cde00c02
  Serguei Katkov authored Jul 25, 2019
  cde00c02
Jul 22, 2019
- [Loop Peeling] Fix the handling of branch weights of peeled off branches. · c6c31da8
  Serguei Katkov authored Jul 22, 2019
  c6c31da8
Jul 15, 2019
- [LoopUnroll+LoopUnswitch] do not transform loops containing callbr · c4f245b4
  Nick Desaulniers authored Jul 15, 2019
  c4f245b4
- [Testing] Add missing "REQUIRES: asserts" · 12400b97
  David Zarzycki authored Jul 15, 2019
  12400b97
- [Loop Peeling] Fix the bug with IDom setting for exit loops · d021ad9f
  Serguei Katkov authored Jul 15, 2019
  d021ad9f
- [Loop Peeling] Enable peeling for loops with multiple exits · 3ed93b46
  Serguei Katkov authored Jul 15, 2019
  3ed93b46
Jun 26, 2019
- [LoopUnroll] Add support for loops with exiting headers and uncond latches. · 4c11b526
  Florian Hahn authored Jun 26, 2019
  4c11b526
Jun 24, 2019
- InstCombine: Preserve nuw when reassociating nuw ops [1/3] · 5a89ba73
  Matt Arsenault authored Jun 24, 2019
  5a89ba73