Commits · f4a09e55a6865ce812d2a24c500d88c340fd4897 · Roger Ferrer / llvm-epi · GitLab

Jun 16, 2017
- [AMDGPU] Testing commit access only, no real change · f9b521fd
  Alfred Huang authored Jun 15, 2017
  f9b521fd
Jun 15, 2017
- DivergencyAnalysis patch for review · 0f9c84cd
  Alexander Timofeev authored Jun 15, 2017
  0f9c84cd
Jun 14, 2017
- [AMDGPU] Remove now dead defaultOffsetS13(). NFCI. · 36559b25
  Davide Italiano authored Jun 13, 2017
  36559b25
Jun 12, 2017
- AMDGPU/GlobalISel: Mark 32-bit G_ADD as legal · ee6e6452
  Tom Stellard authored Jun 12, 2017
  ee6e6452
- AMDGPU: Don't add same implicit use multiple times · 05c26472
  Matt Arsenault authored Jun 12, 2017
  05c26472
- AMDGPU: Teach isLegalAddressingMode about flat offsets · d9b77848
  Matt Arsenault authored Jun 12, 2017
  d9b77848
- AMDGPU: Start selecting flat instruction offsets · db7c6a87
  Matt Arsenault authored Jun 12, 2017
  db7c6a87
- AMDGPU: Verify that flat offsets aren't used pre-GFX9 · 89ad17ce
  Matt Arsenault authored Jun 12, 2017
  89ad17ce
- AMDGPU: Start adding offset fields to flat instructions · fd023141
  Matt Arsenault authored Jun 12, 2017
  fd023141
- Const correctness for TTI::getRegisterBitWidth · c0112ae8
  Daniel Neilson authored Jun 12, 2017
  c0112ae8
Jun 10, 2017
- AMDGPU : Fix ISA Version Definitions. · 7c3e5115
  Wei Ding authored Jun 10, 2017
  7c3e5115
Jun 09, 2017
- [AMDGPU] Add intrinsics for alignbit and alignbyte instructions · 1a61ab81
  Stanislav Mekhanoshin authored Jun 09, 2017
  1a61ab81
- [AMDGPU] Fix for issue in alloca to vector promotion pass · 82618baa
  David Stuttard authored Jun 09, 2017
  82618baa
Jun 08, 2017
- AMDGPU: Work around build special casing .inc files · f1202e65
  Matt Arsenault authored Jun 08, 2017
  f1202e65
- AMDGPU: Use correct register names in inline assembly · 3c7581bb
  Matt Arsenault authored Jun 08, 2017
  3c7581bb
- [AMDGPU] Force qsads instrs to use different dest register than source registers · e5c78323
  Mark Searles authored Jun 08, 2017
  e5c78323
Jun 07, 2017
- [AMDGPU][MC] Corrected error message for s_waitcnt helpers · 5a2f881b
  Dmitry Preobrazhensky authored Jun 07, 2017
  5a2f881b
- AMDGPU/GlobalISel: Mark 32-bit G_SELECT as legal · 2860a428
  Tom Stellard authored Jun 07, 2017
  2860a428
- Move Object format code to lib/BinaryFormat. · 264b5d9e
  Zachary Turner authored Jun 07, 2017
  264b5d9e
Jun 06, 2017
- AMDGPU/NFC: Move amdgpu code object metadata to support · 1e2b8789
  Konstantin Zhuravlyov authored Jun 06, 2017
  1e2b8789
- [AMDGPU] Return correct value from SDWA pass · e4cda741
  Stanislav Mekhanoshin authored Jun 06, 2017
  e4cda741
- AMDGPU/GlobalISel: Mark 32-bit G_ICMP as legal · 8cd60a50
  Tom Stellard authored Jun 06, 2017
  8cd60a50
- Sort the remaining #include lines in include/... and lib/.... · 6bda14b3
  Chandler Carruth authored Jun 06, 2017
  6bda14b3
- [llvm] Remove double semicolons · 5e1697ef
  Mandeep Singh Grang authored Jun 06, 2017
  5e1697ef
Jun 05, 2017
- AMDGPU: Remove deprecated and unused elf definitions · 5b0bf2ff
  Konstantin Zhuravlyov authored Jun 05, 2017
  5b0bf2ff
- [AMDGPU] Fix uninit'ed var (RevisitLoop) · 602ee930
  Mark Searles authored Jun 05, 2017
  602ee930
- [AMDGPU] Fix SIFoldOperands crash with clamp · 286a4225
  Stanislav Mekhanoshin authored Jun 05, 2017
  286a4225
Jun 03, 2017
- [AMDGPU] Untangle SDWA pass from SIShrinkInstructions · 03306604
  Stanislav Mekhanoshin authored Jun 03, 2017
  03306604
- AMDGPU/GlobalISel: Mark 1-bit integer constants as legal · e042412e
  Tom Stellard authored Jun 03, 2017
  e042412e
- [AMDGPU] Preserve operand order in SIFoldOperands · f154b4f5
  Stanislav Mekhanoshin authored Jun 03, 2017
  f154b4f5
- [AMDGPU] V_DIV_FIXUP_F16 is not a commutable operation · ca5d2efe
  Stanislav Mekhanoshin authored Jun 03, 2017
  ca5d2efe
Jun 02, 2017
- AMDGPU: Register AMDGPUAlwaysInline · 746e0657
  Matt Arsenault authored Jun 02, 2017
  746e0657
- AMDGPU: Make auto waitcnt before barrier a feature · be6c0ca5
  Konstantin Zhuravlyov authored Jun 02, 2017
  be6c0ca5
- AMDGPUAnnotateUniformValue should always treat volatile loads as divergent · 3f70b619
  Alexander Timofeev authored Jun 02, 2017
  3f70b619
- [AMDGPU] Turn on the new waitcnt insertion pass. Adjust tests. · 70359ac6
  Mark Searles authored Jun 02, 2017
  70359ac6
Jun 01, 2017
- [AMDGPU] Fix kernel arg segment size for amdgizcl · a618acf9
  Yaxun Liu authored Jun 01, 2017
  a618acf9
- AMDGPU: Remove error on call in AsmPrinter · 3416b8c8
  Matt Arsenault authored Jun 01, 2017
  3416b8c8
- AMDGPU: Set high getCSRFirstUseCost · 50f43e41
  Matt Arsenault authored Jun 01, 2017
  50f43e41
May 31, 2017
- TargetMachine: Indicate whether machine verifier passes. · d6a36ae2
  Matthias Braun authored May 31, 2017
  d6a36ae2
- [AMDGPU] Fix bugs in new waitcnt pass. Add test. · 11d0a040
  Mark Searles authored May 31, 2017
  11d0a040