Commits · f7866fad5415ff8406fe7fa453fed1c4b08f3b92 · Roger Ferrer / llvm-epi · GitLab

Dec 02, 2016
- [AVX-512] Add masked VINSERTF/VINSERTI instructions to load folding tables. · f7866fad
  Craig Topper authored Dec 02, 2016
  f7866fad
- [DWARF] Put linkage-name on abstract origin even when there's a declaration. · dad4907b
  Paul Robinson authored Dec 02, 2016
  dad4907b
- [ThinLTO] Stop importing constant global vars as copies in the backend · 185b4ab6
  Teresa Johnson authored Dec 02, 2016
  185b4ab6
- AMDGPU: Use wider scalar spills for SGPR spilling · c47701c0
  Matt Arsenault authored Dec 02, 2016
  c47701c0
- When instructions are hoisted out of loops by MachineLICM, remove their debug loc. · 42f92a72
  Wolfgang Pieb authored Dec 02, 2016
  42f92a72
- [AArch64] Fold more spilled/refilled COPYs. · 7ffce7be
  Geoff Berry authored Dec 01, 2016
  7ffce7be
- llvm-modextract: Call keep() on the output stream before exiting. · 85c2184a
  Peter Collingbourne authored Dec 01, 2016
  85c2184a
Dec 01, 2016
- [ARM] Fix for 64-bit CAS expansion on ARM32 with -O0 · e2ae4151
  Oleg Ranevskyy authored Dec 01, 2016
  e2ae4151
- Revert "[SLP] Fix for PR6246: vectorization for scalar ops on vector elements." · 704395a2
  Artem Belevich authored Dec 01, 2016
  704395a2
- RegisterCoalscer: Only coalesce complete reserved registers. · 709a4cc2
  Matthias Braun authored Dec 01, 2016
  709a4cc2
- [SelectionDAG] getRawSubclassData should not return HasDebugValue. · 76b913c4
  Chih-Hung Hsieh authored Dec 01, 2016
  76b913c4
- AArch64: fix 128-bit cmpxchg at -O0 (again, again). · 5bb87b67
  Tim Northover authored Dec 01, 2016
  5bb87b67
- [PR29121] Don't fold if it would produce atomic vector loads or stores · 89e92d21
  Philip Reames authored Dec 01, 2016
  89e92d21
- [SLP] Fix for PR6246: vectorization for scalar ops on vector elements. · 2c01af59
  Alexey Bataev authored Dec 01, 2016
  2c01af59
- [tablegen] Delete duplicates from a vector without skipping elements · 47de8391
  Vedant Kumar authored Dec 01, 2016
  47de8391
- Fix a bug with llvm-size and the -m option with multiple files not printing the file names. · 5997c948
  Kevin Enderby authored Dec 01, 2016
  5997c948
- [SLP] Fixed cost model for horizontal reduction. · 62af7252
  Alexey Bataev authored Dec 01, 2016
  62af7252
- [AsmParser] Diagnose empty symbol for .set directive · cf26d563
  Weiming Zhao authored Dec 01, 2016
  cf26d563
- [GVN, OptDiag] Print the interesting instructions involved in missed load-elimination · 4ddb8c01
  Adam Nemet authored Dec 01, 2016
  4ddb8c01
- [GVN, OptDiag] Include the value that is forwarded in load elimination · 8b5fba80
  Adam Nemet authored Dec 01, 2016
  8b5fba80
- [SLP] Additional tests with the cost of vector operations. · fc617690
  Alexey Bataev authored Dec 01, 2016
  fc617690
- Revert "[SLP] Additional tests with the cost of vector operations." · e59a8351
  Alexey Bataev authored Dec 01, 2016
  e59a8351
- [GVN] Basic optimization remark support · 4d2a6e59
  Adam Nemet authored Dec 01, 2016
  4d2a6e59
- [SLP] Additional tests with the cost of vector operations. · 2ff76847
  Alexey Bataev authored Dec 01, 2016
  2ff76847
- [X86][SSE] Classify AND bitmasks as variable shuffle masks · 5fe62360
  Simon Pilgrim authored Dec 01, 2016
  5fe62360
- [X86][SSE] Add support for combining AND bitmasks to shuffles. · 1e4d8709
  Simon Pilgrim authored Dec 01, 2016
  1e4d8709
- [LMT] Restrict nop length to one · 7f6968ed
  Asaf Badouh authored Dec 01, 2016
  7f6968ed
- [X86][SSE] Tidied up filecheck prefixes for uitofp fast-math tests. · 2cff28dd
  Simon Pilgrim authored Dec 01, 2016
  2cff28dd
- [X86][SSE] Add support for combining target shuffles to AND bitmasks. · 55066e56
  Simon Pilgrim authored Dec 01, 2016
  55066e56
- [X86][SSE] Add support for combining ISD::AND with shuffles. · 947650e9
  Simon Pilgrim authored Dec 01, 2016
  947650e9
- [X86][SSE] Added tests showing missed combines of shuffles with ANDs. · ed4ede0c
  Simon Pilgrim authored Dec 01, 2016
  ed4ede0c
- [GVN] When merging blocks update LoopInfo if it's available · feafcd96
  Adam Nemet authored Dec 01, 2016
  feafcd96
- revert r288283 as it causes debug info (line numbers) to be lost in... · b66cb88c
  Kostya Serebryany authored Dec 01, 2016
  b66cb88c
- [WebAssembly] Emit .import_global assembler directives · 7747d703
  Derek Schuff authored Dec 01, 2016
  7747d703
- MCStreamer: Use "cfi" for CFI related temp labels. · 39c3c89c
  Matthias Braun authored Nov 30, 2016
  39c3c89c
- llvm-lto2: Simpler workaround for PR30396. · a5b71649
  Peter Collingbourne authored Nov 30, 2016
  a5b71649
Nov 30, 2016
- Recommit r288212: Emit 'no line' information for interesting 'orphan' instructions. · 37a13ddb
  Paul Robinson authored Nov 30, 2016
  37a13ddb
- Only computeRelativePath() on new members · 5cb34077
  David Callahan authored Nov 30, 2016
  5cb34077
- [AArch64] Refactor LSE support as feature separate from V8.1a support. · 75818bc8
  Joel Jones authored Nov 30, 2016
  75818bc8
- [LoopUnroll] Implement profile-based loop peeling · b151a641
  Michael Kuperstein authored Nov 30, 2016
  b151a641