Commits · dad3d13b993e8394e31dfa2d6f5644357ff941a4 · Roger Ferrer / llvm-epi · GitLab

Mar 19, 2018
- [AMDGPU] adjust tests to be nan-free · dad3d13b
  Sanjay Patel authored Mar 19, 2018
  dad3d13b
- [Power9]Legalize and emit code for quad-precision copySign/abs/nabs/neg/sqrt · ecfede94
  Lei Huang authored Mar 19, 2018
  ecfede94
- [Hexagon] Add a few more lit tests · 461e6691
  Krzysztof Parzyszek authored Mar 19, 2018
  461e6691
- [X86] Remove OUT32rr/OUT8rr/OUT32ri/OUT8ri from Sandybridge scheduler model. · 5e65996f
  Craig Topper authored Mar 19, 2018
  5e65996f
- [X86] Add JCXZ/JECXZ to Sandybridge/Haswell/Broadwell/Skylake scheduler models. · b4c7873f
  Craig Topper authored Mar 19, 2018
  b4c7873f
- [X86] Correct regular expression in Zen scheduler model that was excluding JECXZ instruction. · afabf365
  Craig Topper authored Mar 19, 2018
  afabf365
- [PowerPC][Power9]Legalize and emit code for quad-precision add/div/mul/sub · 6d1596a9
  Lei Huang authored Mar 19, 2018
  6d1596a9
- [PowerPC] Make AddrSpaceCast noop · d9d5bd30
  Nemanja Ivanovic authored Mar 19, 2018
  d9d5bd30
- [X86] Remove sse41 specific code from lowering v16i8 multiply · 259eaa6e
  Craig Topper authored Mar 19, 2018
  259eaa6e
- [X86] Make the multiply and divide itineraries more consistent. · 5ccd8723
  Craig Topper authored Mar 19, 2018
  5ccd8723
- Revert [MachineLICM] This reverts commit rL327856 · 01f414ba
  Zaara Syeda authored Mar 19, 2018
  01f414ba
- [CodeGen] Avoid handling DBG_VALUE in the LivePhysRegs (addUses,removeDefs,stepForward) · 4b54e5fc
  Matt Davis authored Mar 19, 2018
  4b54e5fc
- [InstCombine] canonicalize fcmp+select to fabs · 0ce30867
  Sanjay Patel authored Mar 19, 2018
  0ce30867
- [MachineLICM] Add functions to MachineLICM to hoist invariant stores · ff05e2b0
  Zaara Syeda authored Mar 19, 2018
  ff05e2b0
- [X86] Generalize schedule classes to support multiple stages · 30c38c38
  Simon Pilgrim authored Mar 19, 2018
  30c38c38
- [x86] put nops into the WriteNop class and customize for Jaguar · 05daae75
  Sanjay Patel authored Mar 19, 2018
  05daae75
- TableGen: Explicitly forbid self-references to field members · 2ad19016
  Nicolai Haehnle authored Mar 19, 2018
  2ad19016
- TableGen: Check the dynamic type of !cast<Rec>(string) · 4186cc7c
  Nicolai Haehnle authored Mar 19, 2018
  4186cc7c
- TableGen: Explicitly test some cases of self-references and !cast errors · 18f1998a
  Nicolai Haehnle authored Mar 19, 2018
  18f1998a
- TableGen: Remove the cast-from-string-to-variable-reference feature · c47fe129
  Nicolai Haehnle authored Mar 19, 2018
  c47fe129
- AMDGPU/GlobalISel: RegBankSelect for basic int ops · fed0a450
  Matt Arsenault authored Mar 19, 2018
  fed0a450
- AMDGPU: Don't leave dead illegal VGPR->SGPR copies · 69932e4d
  Matt Arsenault authored Mar 19, 2018
  69932e4d
- [MergeICmps] Re-land 324317 "Enable the MergeICmps Pass by default." · 6d047b70
  Clement Courbet authored Mar 19, 2018
  6d047b70
- [ARM] Support for v4f16 and v8f16 vectors · d16037d9
  Sjoerd Meijer authored Mar 19, 2018
  d16037d9
- [SystemZ] Bugfix of CC liveness in emitMemMemWrapper (CLC). · a6216ec4
  Jonas Paulsson authored Mar 19, 2018
  a6216ec4
- [RISCV] Peephole optimisation for load/store of global values or constant addresses · 0171a9f4
  Alex Bradbury authored Mar 19, 2018
  0171a9f4
- [SCEV] Re-land: Fix isKnownPredicate · 529f4233
  Serguei Katkov authored Mar 19, 2018
  529f4233
- [X86] Add ADD16i16/ADD32i32/ADD64i32 and similar to the scheduler models to match ADD8i8. · d10ceffa
  Craig Topper authored Mar 19, 2018
  d10ceffa
- [AVR] Lower i128 divisions to runtime library calls · a35ee706
  Dylan McKay authored Mar 19, 2018
  a35ee706
Mar 18, 2018
- [X86][Btver2] Fix crc32 schedule costs · 203876f1
  Simon Pilgrim authored Mar 18, 2018
  203876f1
- [X86][Btver2] Add crc32 resource tests · 13cd3b09
  Simon Pilgrim authored Mar 18, 2018
  13cd3b09
- [X86][Btver2] FADD/FHADD ymm instructions are double pumped on the JFPA functional pipe · c3db8c7c
  Simon Pilgrim authored Mar 18, 2018
  c3db8c7c
- [X86][Btver2] Float bitwise ymm instructions are double pumped on the JFPX... · 036cc826
  Simon Pilgrim authored Mar 18, 2018
  036cc826
- [X86][Btver2] F16C instructions are performed on the JSTC functional pipe · 87d2f746
  Simon Pilgrim authored Mar 18, 2018
  87d2f746
- [LICM] Salvage DI from dying Instructions · 3a589103
  Anastasis Grammenos authored Mar 18, 2018
  3a589103
- [InstCombine] peek through unsigned FP casts for zero-equality compares (PR36682) · e6da3063
  Roman Lebedev authored Mar 18, 2018
  e6da3063
- [InstCombine] add nnan requirement for sqrt(x) * sqrt(y) -> sqrt(x*y) · 63b10289
  Sanjay Patel authored Mar 18, 2018
  63b10289
- [InstSimplify] loosen FMF for sqrt(X) * sqrt(X) --> X · 95ec4a4d
  Sanjay Patel authored Mar 18, 2018
  95ec4a4d
- [X86][Btver2] SSE4A EXTRQ/INSERTQ instructions are performed on the JVALU0/JVALU1 functional pipes · 40f6d6ad
  Simon Pilgrim authored Mar 18, 2018
  40f6d6ad
- [X86][Btver2] Modelled float bitwise instructions as being performed on the... · e16790b1
  Simon Pilgrim authored Mar 18, 2018
  e16790b1