Commits · e8df879948f9da611d88224d87af177ee67be2a7 · Roger Ferrer / llvm-epi · GitLab

Aug 22, 2015
- AMDGPU: Improve accuracy of instruction rates for some FP instructions · e8df8799
  Matt Arsenault authored Aug 22, 2015
  e8df8799
- Improve the determinism of MergeFunctions · 057292a7
  JF Bastien authored Aug 21, 2015
  057292a7
- [LAA] Hold bounds via ValueHandles during SCEV expansion · 4e533ef7
  Adam Nemet authored Aug 21, 2015
  4e533ef7
- AMDGPU/SI: Better handle s_wait insertion · bd8a0856
  Tom Stellard authored Aug 21, 2015
  bd8a0856
- Re-apply r245635, "[InstCombine] Transform A & (L - 1) u< L --> L != 0" · c86c162a
  Sanjoy Das authored Aug 21, 2015
  c86c162a
Aug 21, 2015
- MIR Serialization: Serialize the pointer IR expression values in the machine · c1136ef3
  Alex Lorenz authored Aug 21, 2015
  c1136ef3
- [ARM] Fix MachO CPU Subtype selection · 366dd9fd
  Vedant Kumar authored Aug 21, 2015
  366dd9fd
- [PowerPC] PPCVSXFMAMutate should not segfault on undef input registers · ff9639d6
  Hal Finkel authored Aug 21, 2015
  ff9639d6
- Line endings fix. · 76b91d90
  Simon Pilgrim authored Aug 21, 2015
  76b91d90
- [x86] enable machine combiner reassociations for 256-bit vector min/max · f0bc07f7
  Sanjay Patel authored Aug 21, 2015
  f0bc07f7
- remove 'FeatureSlowUAMem' from AMD CPUs based on 10H micro-arch or later · dddad102
  Sanjay Patel authored Aug 21, 2015
  dddad102
- [x86] enable machine combiner reassociations for 128-bit vector min/max · cf942fa9
  Sanjay Patel authored Aug 21, 2015
  cf942fa9
- save some testing time; get rid of the non-SSE chips in this test · c9f4aa6b
  Sanjay Patel authored Aug 21, 2015
  c9f4aa6b
- add a test case to check the fast-unaligned-mem attribute per CPU · 44d7bef0
  Sanjay Patel authored Aug 21, 2015
  44d7bef0
- [DAGCombiner] Fold together mul and shl when both are by a constant · eab960c4
  John Brawn authored Aug 21, 2015
  eab960c4
- Revert r245635, "[InstCombine] Transform A & (L - 1) u< L --> L != 0" · 6a623281
  NAKAMURA Takumi authored Aug 21, 2015
  6a623281
- [Sparc] Support user-specified stack object overalignment. · 667395f3
  James Y Knight authored Aug 21, 2015
  667395f3
- TransformUtils: Introduce module splitter. · 1dc6a8d1
  Peter Collingbourne authored Aug 21, 2015
  1dc6a8d1
- AArch64: Fix testcase of r245640 · 25fd09a7
  Matthias Braun authored Aug 21, 2015
  25fd09a7
- [SLP] Add one more test case for propagating 'nontemporal' attributes. · 6002295c
  Michael Zolotukhin authored Aug 21, 2015
  6002295c
- delete more dead code from this testcase. · 1d8741bc
  Adrian Prantl authored Aug 21, 2015
  1d8741bc
- Further reduce the IR in this testcase based on a further reduction · 62222edb
  Adrian Prantl authored Aug 20, 2015
  62222edb
- AArch64: Fix cmp;ccmp ordering · 46e56398
  Matthias Braun authored Aug 20, 2015
  46e56398
- AArch64: Do not create CCMP on multiple users. · 266204b7
  Matthias Braun authored Aug 20, 2015
  266204b7
- [InstSimplify] add nuw %x, C2 must be at least C2 · 2df38cd0
  David Majnemer authored Aug 20, 2015
  2df38cd0
- [InstCombine] Transform A & (L - 1) u< L --> L != 0 · e472d8a5
  Sanjoy Das authored Aug 20, 2015
  e472d8a5
- [SLP] Propagate 'nontemporal' attribute into vectorized instructions. · 51b00e6d
  Michael Zolotukhin authored Aug 20, 2015
  51b00e6d
- [LoopVectorize] Propagate 'nontemporal' attribute into vectorized instructions. · 2a3d99fe
  Michael Zolotukhin authored Aug 20, 2015
  2a3d99fe
Aug 20, 2015
- [X86] Look for scalar through one bitcast when lowering to VBROADCAST. · 0cdc7719
  Ahmed Bougacha authored Aug 20, 2015
  0cdc7719
- [X86] Add some broadcast-from-memory tests. · 69a17acb
  Ahmed Bougacha authored Aug 20, 2015
  69a17acb
- [NVPTX] truncating 64-bit to 32-bit is free · ca3ef11a
  Jingyue Wu authored Aug 20, 2015
  ca3ef11a
- [X86] Replace avx2 broadcast intrinsics with native IR. · 1a498705
  Ahmed Bougacha authored Aug 20, 2015
  1a498705
- Fix a bug that caused SimplifyCFG to drop DebugLocs. · baf90fc2
  Adrian Prantl authored Aug 20, 2015
  baf90fc2
- Fix a debug location handling bug in GVN. · a317cd25
  Adrian Prantl authored Aug 20, 2015
  a317cd25
- Fix symbol value computation when part of the expression is weak. · c30c7c49
  Rafael Espindola authored Aug 20, 2015
  c30c7c49
- [Sparc]: correct the 'set' synthetic instruction · 58195a2d
  Douglas Katzman authored Aug 20, 2015
  58195a2d
- Optimize bitwise even/odd test (-x&1 -> x&1) to not use negation. · ccf59731
  Balaram Makam authored Aug 20, 2015
  ccf59731
- [mips][microMIPS] Add microMIPS32r6 and microMIPS64r6 tests for existing... · 56585d51
  Zoran Jovanovic authored Aug 20, 2015
  56585d51
- [X86] Fix FBLD and FBSTP · bce1ab67
  Marina Yatsina authored Aug 20, 2015
  bce1ab67
- [X86] Fix bug in COMISD and COMISS definition in td files · 7a4e1ba7
  Marina Yatsina authored Aug 20, 2015
  7a4e1ba7