Commits · e7d8ed32f9f98d02801c2a0a2dd094689e830a21 · Roger Ferrer / llvm-epi · GitLab

Dec 22, 2016
- AMDGPU: Swap order of operands in fadd/fsub combine · e7d8ed32
  Matt Arsenault authored Dec 22, 2016
  e7d8ed32
- AMDGPU: Check fast math flags in fadd/fsub combines · 46e6b7ad
  Matt Arsenault authored Dec 22, 2016
  46e6b7ad
- AMDGPU: Form more FMAs if fusion is allowed · 770ec868
  Matt Arsenault authored Dec 22, 2016
  770ec868
- AMDGPU: Enable some f32 fadd/fsub combines for f16 · ef82ad94
  Matt Arsenault authored Dec 22, 2016
  ef82ad94
- AMDGPU: Implement isFMAFasterThanFMulAndFAdd for f16 · 9e22bc2c
  Matt Arsenault authored Dec 22, 2016
  9e22bc2c
- AMDGPU: setcc test cleanup · 2920f624
  Matt Arsenault authored Dec 22, 2016
  2920f624
- AMDGPU: Allow rcp and rsq usage with f16 · cdff21b1
  Matt Arsenault authored Dec 22, 2016
  cdff21b1
- AMDGPU: Custom lower f16 fdiv · 4052a576
  Matt Arsenault authored Dec 22, 2016
  4052a576
- AMDGPU: Implement f16 fcanonicalize · ce84130f
  Matt Arsenault authored Dec 22, 2016
  ce84130f
- [AArch64] Correct the check of signed 9-bit imm in getIndexedAddressParts(). · 9ac20a1e
  Haicheng Wu authored Dec 22, 2016
  9ac20a1e
- [NVVMIntrRange] Only set range metadata if none is already present · 5fa7d48b
  David Majnemer authored Dec 22, 2016
  5fa7d48b
- Renumber testcase metadata nodes after r290153. · 1eadba1c
  Adrian Prantl authored Dec 22, 2016
  1eadba1c
- [LLParser] Make the line field of DIMacro(File) optional. · 58c19106
  Adrian Prantl authored Dec 22, 2016
  58c19106
- Legalize metadata in legacy testcases · ec9ebba7
  Adrian Prantl authored Dec 21, 2016
  ec9ebba7
- Legalize metadata in legacy testcases · 762e4b72
  Adrian Prantl authored Dec 21, 2016
  762e4b72
- Legalize metadata in legacy testcases · aad5df48
  Adrian Prantl authored Dec 21, 2016
  aad5df48
- Legalize metadata in legacy testcases · b767f312
  Adrian Prantl authored Dec 21, 2016
  b767f312
- IR: Function summary representation for type tests. · 1b4137a7
  Peter Collingbourne authored Dec 21, 2016
  1b4137a7
Dec 21, 2016
- [sancov] hash prefix results in huge merge files, use shorter prefix · 987f6420
  Mike Aizatsky authored Dec 21, 2016
  987f6420
- [X86][SSE] Improve lowering of vXi64 multiplies · 081abbb1
  Simon Pilgrim authored Dec 21, 2016
  081abbb1
- Revert "[InstCombine] New opportunities for FoldAndOfICmp and FoldXorOfICmp" · b0761a0c
  David Majnemer authored Dec 21, 2016
  b0761a0c
- [LLParser] Parse vector GEP constant expression correctly · 88f15eed
  Michael Kuperstein authored Dec 21, 2016
  88f15eed
- [ConstantFolding] Fix vector GEPs harder · dd92c786
  Michael Kuperstein authored Dec 21, 2016
  dd92c786
- Added a template for building target specific memory node in DAG. · 7c7bf1b4
  Elena Demikhovsky authored Dec 21, 2016
  7c7bf1b4
- [X86] Vectorcall Calling Convention - Adding CodeGen Complete Support · cecc4af4
  Oren Ben Simhon authored Dec 21, 2016
  cecc4af4
- [X86] Vectorcall Calling Convention - Adding CodeGen Complete Support · 3b951570
  Oren Ben Simhon authored Dec 21, 2016
  3b951570
- [LDist] Match behavior between invoking via optimization pipeline or opt -loop-distribute · 32e6a34c
  Adam Nemet authored Dec 21, 2016
  32e6a34c
- remove pretty-print test that requires debug · 1857800c
  Sebastian Pop authored Dec 21, 2016
  1857800c
- machine combiner: fix pretty printer · 77794843
  Sebastian Pop authored Dec 21, 2016
  77794843
- [Analysis] Centralize objectsize lowering logic. · 3f08914e
  George Burgess IV authored Dec 20, 2016
  3f08914e
Dec 20, 2016
- Revert "[ObjectYAML] Support for DWARF debug_info section" · abecaa2f
  Chris Bieneman authored Dec 20, 2016
  abecaa2f
- [ObjectYAML] Support for DWARF debug_info section · ffc4aef5
  Chris Bieneman authored Dec 20, 2016
  ffc4aef5
- IR: Eliminate non-determinism in the module summary analysis. · 0c30f089
  Peter Collingbourne authored Dec 20, 2016
  0c30f089
- [ARM] Implement isExtractSubvectorCheap. · d03df814
  Eli Friedman authored Dec 20, 2016
  d03df814
- [ARM] Generate checks for shuffle tests using update_llc_test_checks.py. · 7ce3d799
  Eli Friedman authored Dec 20, 2016
  7ce3d799
- AMDGPU: Allow 16-bit types in inline asm constraints · 9e910142
  Matt Arsenault authored Dec 20, 2016
  9e910142
- AMDGPU: Run fp combine tests on VI · d1ceffcd
  Matt Arsenault authored Dec 20, 2016
  d1ceffcd
- AMDGPU: Don't add same instruction multiple times to worklist · 4c1e9ec0
  Matt Arsenault authored Dec 20, 2016
  4c1e9ec0
- AMDGPU/SI: Add a MachineMemOperand when lowering llvm.amdgcn.buffer.load.* · 6f9ef14b
  Tom Stellard authored Dec 20, 2016
  6f9ef14b
- AMDGPU/SI: Add a MachineMemOperand to MIMG instructions · 244891d1
  Tom Stellard authored Dec 20, 2016
  244891d1