Commits · 53946bf8c68da7661b065e16b07ff085e9641383 · Roger Ferrer / llvm-epi · GitLab

Dec 15, 2015
- [X86] MOVPC32r should only emit CFI adjustments when needed · 53946bf8
  Michael Kuperstein authored Dec 15, 2015
  53946bf8
- [SimplifyCFG] allow speculation of exactly one expensive instruction (PR24818) · 38a02262
  Sanjay Patel authored Dec 15, 2015
  38a02262
- [llvm-profdata] Add support for weighted merge of profile data (2nd try) · 7f5b47dd
  Nathan Slingerland authored Dec 15, 2015
  7f5b47dd
- AMDGPU: mark ldexp LibCalls as unavailable · 78fd4f08
  Nicolai Hahnle authored Dec 15, 2015
  78fd4f08
- [X86] Smaller code for materializing 32-bit 1 and -1 constants · 08d5905b
  Hans Wennborg authored Dec 15, 2015
  08d5905b
- [Hexagon] Preprocess mapped instructions before lowering to MC · 372bd808
  Krzysztof Parzyszek authored Dec 15, 2015
  372bd808
- AMDGPU/SI: Add llvm.amdgcn.mbcnt.* intrinsics · 43f52df0
  Tom Stellard authored Dec 15, 2015
  43f52df0
- AMDGPU/SI: Add llvm.amdgcn.v.interp.p[12] intrinsics · ad7d03da
  Tom Stellard authored Dec 15, 2015
  ad7d03da
- Bitcasts between FP and INT values using direct moves · 8922476b
  Nemanja Ivanovic authored Dec 15, 2015
  8922476b
- Do not try to use i8 and i16 versions of FP_TO_U/SINT soft float library calls · 801ee741
  Michael Kuperstein authored Dec 15, 2015
  801ee741
- Improve the successor list update in TailDuplication.cpp. · 3ba9cf60
  Cong Hou authored Dec 15, 2015
  3ba9cf60
- Type legalizer for masked gather and scatter intrinsics. · 6015f5c8
  Elena Demikhovsky authored Dec 15, 2015
  6015f5c8
- [ShrinkWrapping] Do not choose restore point inside loops. · b82786e0
  Quentin Colombet authored Dec 15, 2015
  b82786e0
- [WebAssembly] Remove .import printing. · dcba3381
  Dan Gohman authored Dec 15, 2015
  dcba3381
- WebAssembly: test global array indexing · 65f0a71f
  JF Bastien authored Dec 15, 2015
  65f0a71f
- Instcombine: destructor loads of structs that do not contains padding · 1c131b37
  Mehdi Amini authored Dec 15, 2015
  1c131b37
- [llvm-readobj] s/FunctionName/LinkageName/ for codeview dumping · 7c0c0c05
  Reid Kleckner authored Dec 15, 2015
  7c0c0c05
- [PGO] make profile prefix even shorter and more readable · c7018a25
  Xinliang David Li authored Dec 15, 2015
  c7018a25
- [X86] Add relaxtion logic for SBB instructions. · 25b43f36
  Quentin Colombet authored Dec 15, 2015
  25b43f36
- [PGO] Shorten profile symbol prefixes · 08127479
  Xinliang David Li authored Dec 14, 2015
  08127479
- Use diagnostic handler in the LLVMContext · 9d2bfc48
  Rafael Espindola authored Dec 14, 2015
  9d2bfc48
- [X86] Add relaxtion logic for ADC instructions. · 2cb8a51c
  Quentin Colombet authored Dec 14, 2015
  2cb8a51c
Dec 14, 2015
- [WebAssembly] Add type prefixes to call instructions · c7c04454
  Dan Gohman authored Dec 14, 2015
  c7c04454
- [WebAssembly] Implement a new algorithm for placing BLOCK markers · 8fe7e86b
  Dan Gohman authored Dec 14, 2015
  8fe7e86b
- Revert "Don't create unnecessary PHIs" · db9a91e3
  Reid Kleckner authored Dec 14, 2015
  db9a91e3
- [X86] Part 2 to fix x86-64 fp128 calling convention. · 7993e18e
  Chih-Hung Hsieh authored Dec 14, 2015
  7993e18e
- add fast-math-flags to 'call' instructions (PR21290) · fa54aced
  Sanjay Patel authored Dec 14, 2015
  fa54aced
- Start implementing FDE dumping when printing the eh_frame. · 52abc5f9
  Pete Cooper authored Dec 14, 2015
  52abc5f9
- Print the eh_frame section in MachoDump. · 23bfa7e9
  Pete Cooper authored Dec 14, 2015
  23bfa7e9
- [ConstantFold] Fix bitcast to gep constant folding transform. · 59be1d65
  David Majnemer authored Dec 14, 2015
  59be1d65
- [MergeFunctions] Use II instead of CI for InvokeInst; NFC · adfec011
  Sanjoy Das authored Dec 14, 2015
  adfec011
- Teach MergeFunctions about operand bundles · 2a74eb00
  Sanjoy Das authored Dec 14, 2015
  2a74eb00
- Teach haveSameSpecialState about operand bundles · 2de4d0aa
  Sanjoy Das authored Dec 14, 2015
  2de4d0aa
- [PGO] Value profiling text format reader/writer support · e3bf4fd3
  Xinliang David Li authored Dec 14, 2015
  e3bf4fd3
- [IR] Remove terminatepad · bbfc7219
  David Majnemer authored Dec 14, 2015
  bbfc7219
- FastISel needs to remove dead code when it bails out. · accc3e03
  Paul Robinson authored Dec 14, 2015
  accc3e03
- [Power PC] llvm soft float support for ppc32 · 280f7101
  Petar Jovanovic authored Dec 14, 2015
  280f7101
- AMDGPU: Use generic bitreverse intrinsic · d079285e
  Matt Arsenault authored Dec 14, 2015
  d079285e
- AMDGPU: Fix splitting vector loads with existing offsets · 52a52a56
  Matt Arsenault authored Dec 14, 2015
  52a52a56
- [InstCombine] fold trunc ([lshr] (bitcast vector) ) --> extractelement (PR25543) · f727e387
  Sanjay Patel authored Dec 14, 2015
  f727e387