Commits · f259efde4730f1c8ba33f6a16f9b93c1e01165f5 · Roger Ferrer / llvm-epi · GitLab

Aug 17, 2010
- PHI elimination should not break back edge. It can cause some significant code... · f259efde
  Evan Cheng authored Aug 17, 2010
  f259efde
- Add a testcase for svn 111208. · eee4824f
  Bob Wilson authored Aug 16, 2010
  eee4824f
- Generalize a pattern for PKHTB: an SRL of 16-31 bits will guarantee · 804f6159
  Bob Wilson authored Aug 16, 2010
  804f6159
- Convert test to FileCheck. · 3fd1e0dc
  Bob Wilson authored Aug 16, 2010
  3fd1e0dc
Aug 16, 2010
- Convert a test to use FileCheck. · 8f553757
  Bob Wilson authored Aug 16, 2010
  8f553757
- Test expects SSE, give him SSE. · cbc55d9d
  Benjamin Kramer authored Aug 15, 2010
  cbc55d9d
Aug 15, 2010
- Restore arch on these test, they fail on arm. · 4566466b
  Benjamin Kramer authored Aug 15, 2010
  4566466b
- Mark as XFAIL on darwin 8. PR 7886. · 339423c4
  Dale Johannesen authored Aug 15, 2010
  339423c4
Aug 14, 2010
- Temporarily disable tail calls on ARM to work around some linker problems. · 3c9ed76b
  Bob Wilson authored Aug 13, 2010
  3c9ed76b
Aug 13, 2010
- Revert 110491. While not wrong, it was based on a · 8d3c89e7
  Dale Johannesen authored Aug 13, 2010
  8d3c89e7
Aug 12, 2010
- - Teach SSEDomainFix to switch between different levels of AVX instructions.... · 7f704b31
  Bruno Cardoso Lopes authored Aug 12, 2010
  7f704b31
- Begin to support some vector operations for AVX 256-bit intructions. The long · 7306c868
  Bruno Cardoso Lopes authored Aug 12, 2010
  7306c868
- This is x86 only test. · 48595bf2
  Devang Patel authored Aug 12, 2010
  48595bf2
Aug 11, 2010
- Add testcases for all AVX 256-bit intrinsics added in the last couple days · 1675ee7a
  Bruno Cardoso Lopes authored Aug 11, 2010
  1675ee7a
- Reapply r109881 using a more strict command line for llc. · 29c8818a
  Bruno Cardoso Lopes authored Aug 11, 2010
  29c8818a
- fix silly typo · a5f923b1
  Jim Grosbach authored Aug 11, 2010
  a5f923b1
- Add a target triple, as the runtime library invocation varies a bit by · 2bf8bd1e
  Jim Grosbach authored Aug 11, 2010
  2bf8bd1e
- Fix test and re-enable it. · b0276814
  Evan Cheng authored Aug 11, 2010
  b0276814
- Temporarily disable some failing tests, until they can be · 4df41148
  Dan Gohman authored Aug 11, 2010
  4df41148
- cortex m4 has floating point support, but only single precision. · 4d5dc3e7
  Jim Grosbach authored Aug 11, 2010
  4d5dc3e7
- Temporarily disable some failing tests, until they can be · f3d783a6
  Dan Gohman authored Aug 11, 2010
  f3d783a6
- Consider this code snippet: · 6a981314
  Bill Wendling authored Aug 11, 2010
  6a981314
- Report error if codegen tries to instantiate a ARM target when the cpu does... · 5190f092
  Evan Cheng authored Aug 11, 2010
  5190f092
- Add ARM Archv6M and let it implies FeatureDB (having dmb, etc.) · 40921a4e
  Evan Cheng authored Aug 11, 2010
  40921a4e
- Add Cortex-M0 support. It's a ARMv6m device (no ARM mode) with some 32-bit · 49e02fc4
  Evan Cheng authored Aug 11, 2010
  49e02fc4
- - Add subtarget feature -mattr=+db which determine whether an ARM cpu has the · 6e809de9
  Evan Cheng authored Aug 11, 2010
  6e809de9
- Update test to match output of optimize compares for ARM. · 79937dfc
  Bill Wendling authored Aug 11, 2010
  79937dfc
- The optimize comparisons pass removes the "cmp" instruction this is checking for. · 871d4e11
  Bill Wendling authored Aug 10, 2010
  871d4e11
Aug 10, 2010
- Re-apply r110655 with fixes. Epilogue must restore sp from fp if the function... · 3f251fb2
  Evan Cheng authored Aug 10, 2010
  3f251fb2
- Revert r110655, "Fix ARM hasFP() semantics. It should return true whenever FP · 0dd47bfc
  Daniel Dunbar authored Aug 10, 2010
  0dd47bfc
- Fix test for more architectures. Patch by Tobias Grosser. · 5730846c
  Jakob Stoklund Olesen authored Aug 10, 2010
  5730846c
- Fix failing testcase. · fedeff80
  Tobias Grosser authored Aug 10, 2010
  fedeff80
- Handle TAG_constant for integers. · b219746c
  Devang Patel authored Aug 10, 2010
  b219746c
- Fix ARM hasFP() semantics. It should return true whenever FP register is · 8d5d1c13
  Evan Cheng authored Aug 10, 2010
  8d5d1c13
Aug 09, 2010
- Have SPU handle halfvec stores aligned by 8 bytes. · 999da1f3
  Kalle Raiskila authored Aug 09, 2010
  999da1f3
Aug 07, 2010
- Use sdmem and sse_load_f64 (etc.) for the vector · a3bd31a9
  Dale Johannesen authored Aug 07, 2010
  a3bd31a9
Aug 06, 2010
- Fix eabi calling convention when a 64 bit value shadows r3. · 027d5bcf
  Rafael Espindola authored Aug 06, 2010
  027d5bcf
- Add an option to always emit realignment code for a particular module. · e1fb772a
  Eric Christopher authored Aug 05, 2010
  e1fb772a
Aug 05, 2010
- Move x86 specific tests into test/CodeGen/X86. · cc3f3b34
  Devang Patel authored Aug 05, 2010
  cc3f3b34
- Move x86-specific tests out of test/Transforms/LoopStrengthReduce and · c53ee449
  Dan Gohman authored Aug 05, 2010
  c53ee449