Commits · e4c8b9b78cb2995576c52d30875bfe4dbbdee7e3 · Roger Ferrer / llvm-epi · GitLab

Apr 07, 2017
- Use PMADDWD to expand reduction in a loop · 58fa7244
  Dehao Chen authored Apr 07, 2017
  58fa7244
- [GlobalISel] implement narrowing for G_CONSTANT. · 2953788c
  Igor Breger authored Apr 07, 2017
  2953788c
- [mips][msa] Fix generation of bm(n)zi and bins[lr]i instructions · bc54eb89
  Petar Jovanovic authored Apr 07, 2017
  bc54eb89
- [SelectionDAG] Enable target specific vector scalarization of calls and returns · 6470ff0b
  Simon Dardis authored Apr 07, 2017
  6470ff0b
- [SystemZ] Check for presence of vector support in SystemZISelLowering · cad72efe
  Jonas Paulsson authored Apr 07, 2017
  cad72efe
- [ARM] GlobalISel: Test hard float properly · fed80723
  Diana Picus authored Apr 07, 2017
  fed80723
- [AMDGPU] Move SiShrinkInstruction and SDWAPeephole to SSAOptimization passes · 6e79529d
  Sam Kolton authored Apr 07, 2017
  6e79529d
- [ARM] GlobalISel: Support frem for 64-bit values · 3c608448
  Diana Picus authored Apr 07, 2017
  3c608448
- [ARM] GlobalISel: Support frem for 32-bit values · a5bab61a
  Diana Picus authored Apr 07, 2017
  a5bab61a
- AMDGPU/GFX9: Fix shared and private aperture queries · 4b3847e8
  Konstantin Zhuravlyov authored Apr 06, 2017
  4b3847e8
- Turn on -addr-sink-using-gep by default. · 5fba1e53
  Eli Friedman authored Apr 06, 2017
  5fba1e53
- [X86] Revert r299387 due to AVX legalization infinite loop. · 6129887d
  Michael Kuperstein authored Apr 06, 2017
  6129887d
Apr 06, 2017
- AMDGPU: Diagnose illegal SGPR to VGPR copies · 21a43825
  Matt Arsenault authored Apr 06, 2017
  21a43825
- AMDGPU: Replace fp16SrcZerosHighBits with a whitelist · 5cf42718
  Matt Arsenault authored Apr 06, 2017
  5cf42718
- [SelectionDAG] [ARM CodeGen] Fix chain information of LowerMUL · 98240e96
  Huihui Zhang authored Apr 06, 2017
  98240e96
- Revert "[ARM] Add Kryo to available targets" · 5e7059b7
  Yi Kong authored Apr 06, 2017
  5e7059b7
- [SDAG] Fix visitAND optimization to deal with vector extract case again. · 974f7c23
  Nirav Dave authored Apr 06, 2017
  974f7c23
- [ARM] Add Kryo to available targets · 2b622b1f
  Yi Kong authored Apr 06, 2017
  2b622b1f
- [AMDGPU] Eliminate barrier if workgroup size is not greater than wavefront size · ea57c385
  Stanislav Mekhanoshin authored Apr 06, 2017
  ea57c385
- [AMDGPU] Resubmit SDWA peephole: enable by default · 9fa16960
  Sam Kolton authored Apr 06, 2017
  9fa16960
- [X86][MMX] Test showing failure to create MMX non-temporal store · 77d3c770
  Simon Pilgrim authored Apr 06, 2017
  77d3c770
- [ARM] Remove a dead ADD during the creation of TBBs · 1b4b59a4
  David Green authored Apr 06, 2017
  1b4b59a4
Apr 05, 2017
- Revert r299536. [AMDGPU] SDWA peephole: enable by default. · d4f70c70
  Ivan Krasin authored Apr 05, 2017
  d4f70c70
- [Hexagon] Use -mattr to select HVX mode in a testcase, NFC · 2182b4b7
  Krzysztof Parzyszek authored Apr 05, 2017
  2182b4b7
- [DAGCombine] Support FMF contract in fused multiple-and-sub too · d5ffdd36
  Adam Nemet authored Apr 05, 2017
  d5ffdd36
- [ARM] Try to re-enable MachineBranchProb.ll for ARM/AArch64 · edfeb773
  Renato Golin authored Apr 05, 2017
  edfeb773
- [SystemZ] Prevent Merging Bitcast with non-normal loads · aa65a2be
  Nirav Dave authored Apr 05, 2017
  aa65a2be
- [DAGCombiner] add and use TLI hook to convert and-of-seteq / or-of-setne to... · b2f1621b
  Sanjay Patel authored Apr 05, 2017
  b2f1621b
- [DAGCombiner] Don't make a BUILD_VECTOR with operands of illegal type. · 38a2da92
  Jonas Paulsson authored Apr 05, 2017
  38a2da92
- [AMDGPU] SDWA peephole: enable by default · 34e29784
  Sam Kolton authored Apr 05, 2017
  34e29784
- [X86] Relax assert in broadcast-of-subvector lowering. · ec8b1fb5
  Ahmed Bougacha authored Apr 05, 2017
  ec8b1fb5
- [AArch64] Avoid partial register deps on insertelt of load into lane 0. · d3c03a5d
  Ahmed Bougacha authored Apr 04, 2017
  d3c03a5d
- Change section flag character for SHF_LINK_ORDER to "o". · 12de7b24
  Evgeniy Stepanov authored Apr 04, 2017
  12de7b24
Apr 04, 2017
- [AArch64][Fuchsia] Allow -mcmodel=kernel for --target=aarch64-fuchsia · 9eb0a1e0
  Petr Hosek authored Apr 04, 2017
  9eb0a1e0
- Verifier: Check some amdgpu calling convention restrictions · 33339687
  Matt Arsenault authored Apr 04, 2017
  33339687
- AMDGPU: Remove legacy export intrinsic · 3e90f848
  Matt Arsenault authored Apr 04, 2017
  3e90f848
- AMDGPU: Remove legacy image intrinsics · 236da200
  Matt Arsenault authored Apr 04, 2017
  236da200
- [X86][LLVM] Converting __mm{|256|512}_movm_epi{8|16|32|64} LLVMIR call into generic intrinsics. · 88fb1710
  Michael Zuckerman authored Apr 04, 2017
  88fb1710
- [tablegen][globalisel] Add support for nested instruction matching. · bee5739a
  Daniel Sanders authored Apr 04, 2017
  bee5739a
- [mips] Deal with empty blocks in the mips hazard scheduler · 0a47edb1
  Simon Dardis authored Apr 04, 2017
  0a47edb1