Commits · 40bae766206f079f82f89df86dc008dba77ba761 · Roger Ferrer / llvm-epi-0.8 · GitLab

This is an archived project. Repository and other project resources are read-only.

Oct 03, 2016
- AMDGPU: Fix missing -verify-machineinstrs in test · 40bae766
  Matt Arsenault authored Oct 03, 2016
  40bae766
Oct 01, 2016
- Revert "AMDGPU: Don't use offen if it is 0" · 86eeda8e
  Mehdi Amini authored Oct 01, 2016
  86eeda8e
- AMDGPU: Don't use offen if it is 0 · 3070fdf7
  Matt Arsenault authored Oct 01, 2016
  3070fdf7
Sep 30, 2016
- AMDGPU: Use unsigned compare for eq/ne · 5d8eb25e
  Matt Arsenault authored Sep 30, 2016
  5d8eb25e
Sep 29, 2016
- AMDGPU: Partially fix control flow at -O0 · e6740754
  Matt Arsenault authored Sep 29, 2016
  e6740754
Sep 28, 2016
- [AMDGPU] Promote uniform i16 ops to i32 ops for targets that have 16 bit instructions · e14df4b2
  Konstantin Zhuravlyov authored Sep 28, 2016
  e14df4b2
- Revert "In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled." · e524f508
  Nirav Dave authored Sep 28, 2016
  e524f508
- In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled. · e17e055b
  Nirav Dave authored Sep 28, 2016
  e17e055b
- [DAG] Remove isVectorClearMaskLegal() check from vector_build dagcombine · 3e06eafc
  Michael Kuperstein authored Sep 28, 2016
  3e06eafc
Sep 26, 2016
- AMDGPU/SI: Don't crash on anonymous GlobalValues · 1b9748c6
  Tom Stellard authored Sep 26, 2016
  1b9748c6
Sep 23, 2016
- AMDGPU/SI: Include implicit arguments in kernarg_segment_byte_size · e88bbc34
  Tom Stellard authored Sep 23, 2016
  e88bbc34
Sep 22, 2016
- [DAG] Fix incorrect alignment of ext load. · 9011da3d
  Nirav Dave authored Sep 22, 2016
  9011da3d
Sep 17, 2016
- AMDGPU: Fix broken FrameIndex handling · ac0fc849
  Matt Arsenault authored Sep 17, 2016
  ac0fc849
- AMDGPU: Push bitcasts through build_vector · d99ef114
  Matt Arsenault authored Sep 17, 2016
  d99ef114
- AMDGPU: Use i64 scalar compare instructions · 7b1dc2c9
  Matt Arsenault authored Sep 17, 2016
  7b1dc2c9
- AMDGPU/SI: Fix kernel argument ABI for HSA · 7998db63
  Tom Stellard authored Sep 16, 2016
  7998db63
- AMDGPU: Allow some control flow intrinsics to be CSEd · 6408c913
  Matt Arsenault authored Sep 16, 2016
  6408c913
Sep 16, 2016
- AMDGPU: Refactor kernel argument lowering · bbeb45af
  Tom Stellard authored Sep 16, 2016
  bbeb45af
- AMDGPU: Use SOPK compare instructions · 7ccf6cd1
  Matt Arsenault authored Sep 16, 2016
  7ccf6cd1
- AMDGPU/SI: Add support for triples with the mesa3d operating system · 0b76fc4c
  Tom Stellard authored Sep 16, 2016
  0b76fc4c
Sep 14, 2016
- Revert "AMDGPU: Use SOPK compare instructions" · f40b70fa
  Matt Arsenault authored Sep 14, 2016
  f40b70fa
- AMDGPU: Use SOPK compare instructions · f757c879
  Matt Arsenault authored Sep 14, 2016
  f757c879
- AMDGPU: Support folding FrameIndex operands · 2bc198a3
  Matt Arsenault authored Sep 14, 2016
  2bc198a3
- AMDGPU: Improve splitting 64-bit bit ops by constants · fa5f767a
  Matt Arsenault authored Sep 14, 2016
  fa5f767a
Sep 13, 2016
- AMDGPU: Support commuting a FrameIndex operand · 25dba300
  Matt Arsenault authored Sep 13, 2016
  25dba300
Sep 12, 2016
- AMDGPU: Do not clobber SCC in SIWholeQuadMode · e58e0e3f
  Nicolai Haehnle authored Sep 12, 2016
  e58e0e3f
- llvm/test/CodeGen/AMDGPU/infinite-loop-evergreen.ll REQUIRES +Asserts. · cf6aaa9e
  NAKAMURA Takumi authored Sep 12, 2016
  cf6aaa9e
Sep 10, 2016
- AMDGPU: Fix immediate folding logic when shrinking instructions · 124384f0
  Matt Arsenault authored Sep 09, 2016
  124384f0
- AMDGPU: Run LoadStoreVectorizer pass by default · 0efdd06b
  Matt Arsenault authored Sep 09, 2016
  0efdd06b
Sep 09, 2016
- AMDGPU : Fix mqsad_u32_u8 instruction incorrect data type. · 06f8d394
  Wei Ding authored Sep 09, 2016
  06f8d394
- AMDGPU/SI: Make sure llvm.amdgcn.implicitarg.ptr() is 8-byte aligned for HSA · b2869eb6
  Tom Stellard authored Sep 09, 2016
  b2869eb6
- GlobalISel: move type information to MachineRegisterInfo. · 0f140c76
  Tim Northover authored Sep 09, 2016
  0f140c76
- [AMDGPU] Assembler: rename amd_kernel_code_t asm names according to spec · a2e5c88b
  Sam Kolton authored Sep 09, 2016
  a2e5c88b
Sep 08, 2016
- AMDGPU: Try to commute when selecting s_addk_i32/s_mulk_i32 · be90f70d
  Matt Arsenault authored Sep 08, 2016
  be90f70d
- AMDGPU: Support commuting with immediate in src0 · bbb47da8
  Matt Arsenault authored Sep 08, 2016
  bbb47da8
- [SelectionDAG] Add BUILD_VECTOR support to computeKnownBits and SimplifyDemandedBits · cc7b4b51
  Simon Pilgrim authored Sep 08, 2016
  cc7b4b51
Sep 07, 2016
- AMDGPU: Add hidden kernel arguments to runtime metadata · 63891400
  Yaxun Liu authored Sep 07, 2016
  63891400
Sep 06, 2016
- [AMDGPU] Wave and register controls · 864718c6
  Konstantin Zhuravlyov authored Sep 06, 2016
  864718c6
- [AMDGPU] Wave and register controls · 1d65026c
  Konstantin Zhuravlyov authored Sep 06, 2016
  1d65026c
- AMDGPU : Add XNACK feature to GPUs that support it. · 5e832e86
  Wei Ding authored Sep 06, 2016
  5e832e86