Commits · f2674319018151e00f603ebab34ab2155f669394 · Roger Ferrer / llvm-epi · GitLab

Jun 21, 2018
- AMDGPU: Turn D16 for MIMG instructions into a regular operand · f2674319
  Nicolai Haehnle authored Jun 21, 2018
  f2674319
- Generalize MergeBlockIntoPredecessor. Replace uses of MergeBasicBlockIntoOnlyPred. · dfd14ade
  Alina Sbirlea authored Jun 20, 2018
  dfd14ade
Jun 20, 2018
- Allow binop C1, (select cc, CF, CT) -> select folding · 20279dc0
  Stanislav Mekhanoshin authored Jun 20, 2018
  20279dc0
- AMDGPU: Fix scalar_to_vector for v4i16/v4f16 · 5a4ec812
  Matt Arsenault authored Jun 20, 2018
  5a4ec812
Jun 19, 2018
- Utilize new SDNode flag functionality to expand current support for fadd · 7b993d76
  Michael Berg authored Jun 18, 2018
  7b993d76
Jun 18, 2018
- Shrink interval after moving copy in removePartialRedundancy · 54601732
  Krzysztof Parzyszek authored Jun 18, 2018
  54601732
Jun 16, 2018
- [AMDGPU] setcc (select cc, CT, CF), CF, eq | ne -> xor cc, -1 | cc · 3b11794d
  Stanislav Mekhanoshin authored Jun 16, 2018
  3b11794d
Jun 15, 2018
- Utilize new SDNode flag functionality to expand current support for fdiv · 02d1c6c0
  Michael Berg authored Jun 15, 2018
  02d1c6c0
- AMDGPU: Add combine for short vector extract_vector_elts · 63bc0e3c
  Matt Arsenault authored Jun 15, 2018
  63bc0e3c
- AMDGPU: Make v4i16/v4f16 legal · 02dc7e19
  Matt Arsenault authored Jun 15, 2018
  02dc7e19
- DAG: Fix creating concat_vectors with illegal type · df2f4ef2
  Matt Arsenault authored Jun 15, 2018
  df2f4ef2
- [AMDGPU] Recognize x & ~(-1 << y) pattern. · dec562c8
  Roman Lebedev authored Jun 15, 2018
  dec562c8
- [AMDGPU] Recognize x & ((1 << y) - 1) pattern. · 9c17dad8
  Roman Lebedev authored Jun 15, 2018
  9c17dad8
- [AMDGPU] Recognize x & (-1 >> (32 - y)) pattern. · aa8587d1
  Roman Lebedev authored Jun 15, 2018
  aa8587d1
Jun 14, 2018
- AMDGPU/GlobalISel: Implement select() for @llvm.amdgcn.cvt.pkrtz · a9284735
  Tom Stellard authored Jun 14, 2018
  a9284735
- AMDGPU/GlobalISel: Implement select() for 32-bit G_FADD and G_FMUL · 46bbbc33
  Tom Stellard authored Jun 13, 2018
  46bbbc33
Jun 13, 2018
- [AMDGPU] Corrected computeKnownBits for V_PERM_B32 · 7bec5730
  Stanislav Mekhanoshin authored Jun 13, 2018
  7bec5730
- [AMDGPU] Change enqueue kernel handle type · fb17bf60
  Yaxun Liu authored Jun 13, 2018
  fb17bf60
- Revert "Improve handling of COPY instructions with identical value numbers" · 3e039f86
  Krzysztof Parzyszek authored Jun 13, 2018
  3e039f86
- Improve handling of COPY instructions with identical value numbers · 36b816f8
  Krzysztof Parzyszek authored Jun 13, 2018
  36b816f8
- [AMDGPU] DAG combine to produce V_PERM_B32 · 8fd3c4e4
  Stanislav Mekhanoshin authored Jun 12, 2018
  8fd3c4e4
Jun 12, 2018
- AMDHSA/NFC: Code object v3 updates (additional): · ce25bc3e
  Konstantin Zhuravlyov authored Jun 12, 2018
  ce25bc3e
- AMDHSA: Code object v3 updates · 00f2cb11
  Konstantin Zhuravlyov authored Jun 12, 2018
  00f2cb11
- [AMDGPU] prevent hitting Assertion `isReg() && "Wrong MachineOperand accessor"' · 987f292c
  Mark Searles authored Jun 12, 2018
  987f292c
Jun 11, 2018
- [AMDGPU] Do not consider indirect acces through phi for wave limiter · 7ba3fc73
  Stanislav Mekhanoshin authored Jun 11, 2018
  7ba3fc73
- [NFC][AMDGPU] Add tests for all the various IR patterns equivalent to extracting low bits. · b896c4e8
  Roman Lebedev authored Jun 11, 2018
  b896c4e8
Jun 08, 2018
- [AMDGPU] Inline asm - added i16, half and i128 types support · c9a098b3
  Daniil Fukalov authored Jun 08, 2018
  c9a098b3
- AMDGPU: Error on LDS global address in functions · 6fc37598
  Matt Arsenault authored Jun 08, 2018
  6fc37598
- [AMDGPU] Simplify memory legalizer · a5a7c331
  Tony Tye authored Jun 07, 2018
  a5a7c331
Jun 07, 2018
- AMDGPU: Fix not including v2f64 in SReg_128 · f1c868ef
  Matt Arsenault authored Jun 07, 2018
  f1c868ef
- AMDGPU: Use scalar operations for f16 fabs/fneg patterns · 697300bd
  Matt Arsenault authored Jun 07, 2018
  697300bd
- AMDGPU: Try a lot harder to emit scalar loads · 90083d30
  Matt Arsenault authored Jun 07, 2018
  90083d30
- [AMDGPU] Improve reciprocal handling · df61be70
  Stanislav Mekhanoshin authored Jun 06, 2018
  df61be70
Jun 06, 2018
- AMDGPU: Custom lower v2f16 fneg/fabs with illegal f16 · e9524f1f
  Matt Arsenault authored Jun 06, 2018
  e9524f1f
Jun 05, 2018
- AMDGPU: Preserve metadata when widening loads · 57e541e8
  Matt Arsenault authored Jun 05, 2018
  57e541e8
- AMDGPU: Use more custom insert/extract_vector_elt lowering · 9224c00d
  Matt Arsenault authored Jun 05, 2018
  9224c00d
- DAG: Stop dropping invariant/dereferencable · 191bc715
  Matt Arsenault authored Jun 05, 2018
  191bc715
Jun 04, 2018
- [CodeGen] Always update divergence in SelectionDAG::UpdateNodeOperands · ba81d7f1
  Scott Linder authored Jun 04, 2018
  ba81d7f1
- [AMDGPU][Waitcnt] Fix handling of flat instrs · f0b93f1e
  Mark Searles authored Jun 04, 2018
  f0b93f1e
Jun 01, 2018
- AMDGPU: Switch some half using-tests to use amdhsa · 72a9f52c
  Matt Arsenault authored Jun 01, 2018
  72a9f52c