Commits · f8b55f1bc5aa5278716a77bcd02915e0d1e757ab · Roger Ferrer / llvm-epi · GitLab

Nov 14, 2014
- [Reassociate] Canonicalize constants to RHS operand. · f8b55f1b
  Chad Rosier authored Nov 14, 2014
  f8b55f1b
- Reapply "[dwarfdump] Add support for dumping accelerator tables." · e837ec29
  Frederic Riss authored Nov 14, 2014
  e837ec29
- [AVX512] Add 512b masked integer shift by immediate patterns. · 04400449
  Cameron McInally authored Nov 14, 2014
  04400449
- R600/SI: Start implementing an assembler · 9d7ddd51
  Tom Stellard authored Nov 14, 2014
  9d7ddd51
- [PowerPC] Add VSX builtins for vec_div · 76746929
  Bill Schmidt authored Nov 14, 2014
  76746929
- X86: use getConstant rather than getTargetConstant behind BUILD_VECTOR. · d3be12a6
  Tim Northover authored Nov 14, 2014
  d3be12a6
- Allow the use of functions as typeinfo in landingpad clauses · 283bc2ed
  Reid Kleckner authored Nov 14, 2014
  283bc2ed
- First stage of call lowering for Mips fast-isel · d5c4196c
  Reed Kotler authored Nov 13, 2014
  d5c4196c
- Fix symbol resolution of floating point libc builtins in MCJIT · 9aeb0479
  Reid Kleckner authored Nov 13, 2014
  9aeb0479
- Relax the gcov version.ll test to check '.' instead of '\*' · 29d880bd
  Reid Kleckner authored Nov 13, 2014
  29d880bd
- R600/SI: Fix fmin_legacy / fmax_legacy matching for SI · da59f3de
  Matt Arsenault authored Nov 13, 2014
  da59f3de
Nov 13, 2014
- Revert "[GVN] Perform Scalar PRE on gep indices that feed loads before doing Load PRE." · 8716b585
  Chad Rosier authored Nov 13, 2014
  8716b585
- [x86] Add some tests for specific patterns of lane-flips combined with · 99b261ce
  Chandler Carruth authored Nov 13, 2014
  99b261ce
- [GVN] Perform Scalar PRE on gep indices that feed loads before doing Load PRE. · dd526665
  Chad Rosier authored Nov 13, 2014
  dd526665
- [FastISel][AArch64] Don't bail during simple GEP instruction selection. · 0af310d0
  Juergen Ributzka authored Nov 13, 2014
  0af310d0
- R600/SI: Use s_movk_i32 · 77849929
  Matt Arsenault authored Nov 13, 2014
  77849929
- R600: Fix assert on empty function · 6ef66144
  Matt Arsenault authored Nov 13, 2014
  6ef66144
- R600: Error on initializer for LDS. · cc8d3b87
  Matt Arsenault authored Nov 13, 2014
  cc8d3b87
- R600/SI: Get rid of FCLAMP_SI pseudo · 1cffa4c1
  Matt Arsenault authored Nov 13, 2014
  1cffa4c1
- R600/SI: Allow commuting with src2_modifiers · 581a7a69
  Matt Arsenault authored Nov 13, 2014
  581a7a69
- R600/SI: Allow commuting some 3 op instructions · 95e48668
  Matt Arsenault authored Nov 13, 2014
  95e48668
- ARM: allow constpool entry to be moved to the user's block in all cases. · 631cc9ce
  Tim Northover authored Nov 13, 2014
  631cc9ce
- ARM: avoid duplicating branches during constant islands. · ab85dcc7
  Tim Northover authored Nov 13, 2014
  ab85dcc7
- ARM: add @llvm.arm.space intrinsic for testing ConstantIslands. · 650b0ee5
  Tim Northover authored Nov 13, 2014
  650b0ee5
- AVX-512: SINT_TO_FP cost model and some bugfixes · d5e95b57
  Elena Demikhovsky authored Nov 13, 2014
  d5e95b57
- OCAMLFLAGS can contain =, don't use = with sed · e353eba3
  Hal Finkel authored Nov 13, 2014
  e353eba3
- Revert r219432 - "Revert "[BasicAA] Revert "Revert r218714 - Make better use... · 45ba2c10
  Hal Finkel authored Nov 13, 2014
  45ba2c10
- [x86] Teach the vector shuffle lowering to make a more nuanced decision · fee91883
  Chandler Carruth authored Nov 13, 2014
  fee91883
- llvm-readobj: Print out address table when dumping COFF delay-import table · ffa4cebe
  Rui Ueyama authored Nov 13, 2014
  ffa4cebe
- Add an assert and a test that verify r221709's fix. · 0f7abef2
  Frederic Riss authored Nov 13, 2014
  0f7abef2
- [x86] Don't form overly fragmented blends when splitting and · 253dd39a
  Chandler Carruth authored Nov 13, 2014
  253dd39a
- [CodeGenPrepare] Handle zero extensions in the TypePromotionHelper. · f5485bb0
  Quentin Colombet authored Nov 13, 2014
  f5485bb0
- [FastISel][AArch64] Optimize select when one of the operands is a 'true' or 'false' value. · 957a1454
  Juergen Ributzka authored Nov 13, 2014
  957a1454
- [FastISel][AArch64] Fold the cmp into the select when possible. · 424c5fd1
  Juergen Ributzka authored Nov 13, 2014
  424c5fd1
- [FastISel][AArch64] Extend 'select' lowering to support also i1 to i16. · d1a042ab
  Juergen Ributzka authored Nov 13, 2014
  d1a042ab
- Revert "[dwarfdump] Add support for dumping accelerator tables." · e1f49581
  Frederic Riss authored Nov 13, 2014
  e1f49581
- Teach ScalarEvolution to sharpen range information. · c5676df3
  Sanjoy Das authored Nov 13, 2014
  c5676df3
- Fix emission of Dwarf accelerator table when there are multiple CUs. · 3a6b354b
  Frederic Riss authored Nov 12, 2014
  3a6b354b
- [dwarfdump] Add support for dumping accelerator tables. · 39467276
  Frederic Riss authored Nov 12, 2014
  39467276
Nov 12, 2014
- [CodeGenPrepare][AArch64] Fix a TLI legality check on iPTR to use a lowered instead. · 0788d49a
  Ahmed Bougacha authored Nov 12, 2014
  0788d49a