Commits · f0d81a50bf4022d62297b842072fbcb5d29c345f · Roger Ferrer / llvm-epi · GitLab

Apr 11, 2015
- [PowerPC] Disable part-word atomics on the P7 · 58f5f9c3
  Hal Finkel authored Apr 11, 2015
  58f5f9c3
- Add direct moves to/from VSR and exploit them for FP/INT conversions · c38b5311
  Nemanja Ivanovic authored Apr 11, 2015
  c38b5311
- [PowerPC] Fix PPCLoopPreIncPrep for depth > 1 loops · ffb460fd
  Hal Finkel authored Apr 11, 2015
  ffb460fd
- [CodeGen] Split -enable-global-merge into ARM and AArch64 options. · b96444ef
  Ahmed Bougacha authored Apr 11, 2015
  b96444ef
- [WinEH] Recognize SEH finally block inserted by the frontend · 9405ef0e
  Reid Kleckner authored Apr 10, 2015
  9405ef0e
- [RewriteStatepointsForGC] test case missing from 234657 · 8106a264
  Philip Reames authored Apr 10, 2015
  8106a264
- [RewriteStatepointsForGC] Use an actual liveness algorithm · df1ef08c
  Philip Reames authored Apr 10, 2015
  df1ef08c
- [RewriteStatepointsForGC] Preprocess the IR to remove unreachable blocks and single entry phis · 85b36a81
  Philip Reames authored Apr 10, 2015
  85b36a81
Apr 10, 2015
- [RewriteStatepointsForGC] Limited support for vectors of pointers · 8531d8c4
  Philip Reames authored Apr 10, 2015
  8531d8c4
- [InstCombine][CodeGenPrep] Create llvm.uadd.with.overflow in CGP. · b6c59143
  Sanjoy Das authored Apr 10, 2015
  b6c59143
- Avoid spewing binary to stdout in some filetype=obj tests · aa7a5a3a
  Reid Kleckner authored Apr 10, 2015
  aa7a5a3a
- use update_llc_test_checks.py to tighten checking · 89e28f67
  Sanjay Patel authored Apr 10, 2015
  89e28f67
- [WinEH] Try to make outlining invokes work a little better · 6e48a826
  Reid Kleckner authored Apr 10, 2015
  6e48a826
- [PowerPC] Prefetching should also consider depth > 1 loops · a9fceb80
  Hal Finkel authored Apr 10, 2015
  a9fceb80
- [mips] [IAS] Make the mips-expansions-bad.s test more readable. NFC. · 038a330e
  Toma Tabacu authored Apr 10, 2015
  038a330e
- Divergence analysis for GPU programs · 5da831cc
  Jingyue Wu authored Apr 10, 2015
  5da831cc
- [PowerPC] Don't crash on PPC32 i64 fp_to_uint on modern cores · 93138503
  Hal Finkel authored Apr 10, 2015
  93138503
- [AArch64] Promote f16 operations to f32. · 1ffe7c7d
  Ahmed Bougacha authored Apr 10, 2015
  1ffe7c7d
- Add LLVM support for remaining integer divide and permute instructions from ISA 2.06 · c0904791
  Nemanja Ivanovic authored Apr 09, 2015
  c0904791
Apr 09, 2015
- [CodeGen] Combine concat_vector of trunc'd scalar to scalar_to_vector. · df437377
  Ahmed Bougacha authored Apr 09, 2015
  df437377
- [AArch64][FastISel] Fix integer extend optimization. · bd0c7eb4
  Juergen Ributzka authored Apr 09, 2015
  bd0c7eb4
- Revert "Refactoring and enhancement to FMA combine." · 1c842716
  Rafael Espindola authored Apr 09, 2015
  1c842716
- Refactoring and enhancement to FMA combine. · 53703d08
  Olivier Sallenave authored Apr 09, 2015
  53703d08
- [ARM] support for Cortex-R4/R4F · 5c5e3c5e
  Javed Absar authored Apr 09, 2015
  5c5e3c5e
- [AArch64] Add support for dynamic stack alignment · 17cb8982
  Kristof Beyls authored Apr 09, 2015
  17cb8982
- [AArch64] Remove redundant -march option. Also fix a think-o from r234462. · 522bf13b
  Lang Hames authored Apr 09, 2015
  522bf13b
- Not all triples put _ before function names. Specify a triple to make this test pass on Linux. · d65d3d7b
  Nick Lewycky authored Apr 09, 2015
  d65d3d7b
- [AArch64] Teach AArch64TargetLowering::getOptimalMemOpType to consider alignment · 90333851
  Lang Hames authored Apr 09, 2015
  90333851
- Use option -march instead of -mtriple to avoid overconditionalizing the test. · 4cb3d064
  Akira Hatanaka authored Apr 08, 2015
  4cb3d064
Apr 08, 2015
- Pass -mtriple to llc to appease buildbot. · bda7aa59
  Akira Hatanaka authored Apr 08, 2015
  bda7aa59
- [WinEH] Minor bug fixes. · 67d3c035
  Andrew Kaylor authored Apr 08, 2015
  67d3c035
- [DAGCombine] Fix a bug in MergeConsecutiveStores. · c6fab805
  Akira Hatanaka authored Apr 08, 2015
  c6fab805
- [LoopAccesses] Allow analysis to complete in the presence of uniform stores · ce48250f
  Adam Nemet authored Apr 08, 2015
  ce48250f
- [ARM] make vminnm/vmaxnm work with ?le, ?ge and no-nans-fp-math · 7ad77920
  Scott Douglass authored Apr 08, 2015
  7ad77920
- fixed to test features, not CPU models · 0f712d04
  Sanjay Patel authored Apr 08, 2015
  0f712d04
- [mips] [IAS] Do not generate redundant move when expanding lw/sw with symbol. · c6ce0749
  Toma Tabacu authored Apr 08, 2015
  c6ce0749
- [mips] [IAS] Add support for the BNEZL and BEQZL pseudo-instructions. · 91fc0b3c
  Toma Tabacu authored Apr 08, 2015
  91fc0b3c
- Write the section header in the end. · 7230f80e
  Rafael Espindola authored Apr 08, 2015
  7230f80e
- [ARM][Debug Info] Restore emitting of .cfi_def_cfa_offset for functions without stack frame · 3cc62b37
  Sergey Dmitrouk authored Apr 08, 2015
  3cc62b37
- [mips] [IAS] Remove AssemblerPredicate's from RelocPIC and RelocStatic. · 7567a10c
  Toma Tabacu authored Apr 08, 2015
  7567a10c