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
Aug 16, 2010
- 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 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
Aug 10, 2010
- 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
Aug 07, 2010
- Use sdmem and sse_load_f64 (etc.) for the vector · a3bd31a9
  Dale Johannesen authored Aug 07, 2010
  a3bd31a9
Aug 06, 2010
- 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
- tests: CodeGen/X86/GC tests require X86. · e62e6646
  Daniel Dunbar authored Aug 05, 2010
  e62e6646
- The lower invoke pass needs to have unreachable code elimination run after it · ca1cb136
  Bill Wendling authored Aug 04, 2010
  ca1cb136
- PR7814: Truncates cannot be ignored for signed comparisons. · 39d0f57c
  Eli Friedman authored Aug 04, 2010
  39d0f57c
Aug 04, 2010
- call-imm.ll test case regex fix. Patch by Dimitry Andric! · cba0d06b
  Stuart Hastings authored Aug 04, 2010
  cba0d06b
Aug 03, 2010
- OK, that's it. This test is going away now. But don't worry, I am taking it to a · 011ff9be
  Jakob Stoklund Olesen authored Aug 03, 2010
  011ff9be
Aug 01, 2010
- Revert new AVX intrinsic tests. They are breaking buildbots and Bruno is · 66161f5e
  Bob Wilson authored Jul 31, 2010
  66161f5e
Jul 30, 2010
- A *bunch* of tests for AVX intrinsics · 92941fdb
  Bruno Cardoso Lopes authored Jul 30, 2010
  92941fdb
- Fix for bug reported by Evzen Muller on llvm-commits: make sure to correctly · ffe64c06
  Eli Friedman authored Jul 30, 2010
  ffe64c06
Jul 28, 2010
- Implement a vectorized algorithm for <16 x i8> << <16 x i8> · 53afc8f0
  Nate Begeman authored Jul 28, 2010
  53afc8f0
- ~40% faster vector shl <4 x i32> on SSE 4.1 Larger improvements for smaller... · 269a6da0
  Nate Begeman authored Jul 27, 2010
  269a6da0
Jul 23, 2010
- Use the proper type for shift counts. This fixes a bootstrap error. · 55e24469
  Dan Gohman authored Jul 23, 2010
  55e24469
- DAGCombine (shl (anyext x, c)) to (anyext (shl x, c)) if the high bits · 0818684a
  Dan Gohman authored Jul 23, 2010
  0818684a
Jul 22, 2010
- Custom lower the memory barrier instructions and add support · 9a773826
  Eric Christopher authored Jul 22, 2010
  9a773826
Jul 20, 2010
- Fix SCEV denormalization of expressions where the exit value from · 625fd229
  Dan Gohman authored Jul 20, 2010
  625fd229
- The same problem was being tracked in PR7652. · 2e839de3
  Duncan Sands authored Jul 20, 2010
  2e839de3
- After a custom inserter, in a block which has constant instructions, · b5e918dc
  Dan Gohman authored Jul 19, 2010
  b5e918dc
Jul 19, 2010
- Remove r108639 now that it is handled by InstCombine instead. · 9c271e28
  Owen Anderson authored Jul 19, 2010
  9c271e28
Jul 18, 2010
- Add a testcase for r108639. · 41670a11
  Owen Anderson authored Jul 18, 2010
  41670a11
Jul 17, 2010
- Consider this function: · bf8370ff
  Bill Wendling authored Jul 16, 2010
  bf8370ff
Jul 16, 2010
- Remove the X86::FP_REG_KILL pseudo-instruction and the X86FloatingPointRegKill · c30b4ddc
  Jakob Stoklund Olesen authored Jul 16, 2010
  c30b4ddc
- Add forgotten test case. · b1671271
  Jakob Stoklund Olesen authored Jul 16, 2010
  b1671271
- Use the source-order scheduler instead of the "fast" scheduler at -O0, · 103c4ebe
  Dan Gohman authored Jul 16, 2010
  103c4ebe
- Revert. This isn't the correct way to go. · 4bda1c8e
  Bill Wendling authored Jul 15, 2010
  4bda1c8e