Commits · d92a8216ecb7f8b96fc7ac804b7f495eaa8693d9 · Roger Ferrer / llvm-epi · GitLab

Sep 08, 2008
- xfail · d92a8216
  Devang Patel authored Sep 08, 2008
  d92a8216
- Update the callgraph correctly in ArgumentPromotion. · 3cf7d865
  Duncan Sands authored Sep 08, 2008
  3cf7d865
Sep 06, 2008
- When PruneEH turned an invoke into an ordinary · 95c2a784
  Duncan Sands authored Sep 06, 2008
  95c2a784
Sep 03, 2008
- Don't crash when trying to constant fold a vector with some elements that can't · f023db64
  Nick Lewycky authored Sep 03, 2008
  f023db64
- Check iteration count. · b530f081
  Devang Patel authored Sep 03, 2008
  b530f081
- If all IV uses are extending integer IV then change the type of IV itself, if possible. · 43c5a52e
  Devang Patel authored Sep 02, 2008
  43c5a52e
- respect inline=never and inline=always notes. · bfa535af
  Devang Patel authored Sep 02, 2008
  bfa535af
Aug 26, 2008
- If IV is used in a int-to-float cast inside the loop then try to eliminate the cast operation. · 4310d398
  Devang Patel authored Aug 26, 2008
  4310d398
Aug 23, 2008
- Fix PR2423 by checking all indices for out of range access, not only · 3f972c91
  Chris Lattner authored Aug 23, 2008
  3f972c91
Aug 21, 2008
- Revert r54876 r54877 r54906 and r54907. Evan found that these caused a 20% · 99f45581
  Nick Lewycky authored Aug 21, 2008
  99f45581
Aug 18, 2008
- XFAIL this test for now. · fe18a8d9
  Bill Wendling authored Aug 18, 2008
  fe18a8d9
Aug 17, 2008
- Consider the case where xor by -1 and xor by 128 have been combined already to · 53b44029
  Nick Lewycky authored Aug 17, 2008
  53b44029
- Didn't mean to change this. · 8ec334f4
  Evan Cheng authored Aug 17, 2008
  8ec334f4
- Fix a (u)comiss intrinsic lowering bug. It was using anyext which can return... · ab35bfdf
  Evan Cheng authored Aug 17, 2008
  ab35bfdf
- Xor'ing both sides of icmp by sign-bit is equivalent to swapping signedness of · 18f50b26
  Nick Lewycky authored Aug 17, 2008
  18f50b26
Aug 16, 2008
- Remove GCSE and LoadVN from the testsuite. · 2a6adfa4
  Owen Anderson authored Aug 16, 2008
  2a6adfa4
Aug 15, 2008
- Reapply 54786. Add overflow and number of mantissa bits checks. · f2a03d5a
  Devang Patel authored Aug 15, 2008
  f2a03d5a
- Revert 54786. It's not checking for overflows, etc. · 86834d29
  Evan Cheng authored Aug 15, 2008
  86834d29
Aug 14, 2008
- If IV is used in a int-to-float cast inside the loop then try to eliminate the cast opeation. · 054a833d
  Devang Patel authored Aug 14, 2008
  054a833d
- Fix a bogus srem rule - a negative value srem'd by a power-of-2 · 6134fbcc
  Dan Gohman authored Aug 13, 2008
  6134fbcc
Aug 13, 2008
- Fix SCCP's handling of struct value loads and stores. SCCP doesn't · 8ded5d58
  Dan Gohman authored Aug 13, 2008
  8ded5d58
- Check sign to detect overflow before changing compare stride. · 97387e66
  Devang Patel authored Aug 13, 2008
  97387e66
Aug 12, 2008
- Implement support for simplifying vector comparisons by 0.0 and 1.0 like we · 2aa0ff27
  Chris Lattner authored Aug 11, 2008
  2aa0ff27
Aug 07, 2008
- Add a basic test for the SRETPromotion pass. · d705b2be
  Matthijs Kooijman authored Aug 07, 2008
  d705b2be
- Move two tests from SRETPromotion to Inline, since they only call opt -inline. · 0620096c
  Matthijs Kooijman authored Aug 07, 2008
  0620096c
Aug 06, 2008
- Fix a shufflevector instcombine that was emitting invalid masks indices · ac22cfca
  Dan Gohman authored Aug 06, 2008
  ac22cfca
- PR2535, not PR2355. · 2bd97afb
  Evan Cheng authored Aug 06, 2008
  2bd97afb
- Fix PR2355: bug in ChangeCompareStride. When the loop termination compare is... · 907dc2bc
  Evan Cheng authored Aug 06, 2008
  907dc2bc
- optimize a common idiom generated by clang for bitfield access, PR2638. · f5b353c1
  Chris Lattner authored Aug 06, 2008
  f5b353c1
- Zap sitofp/fptoui pairs. In all cases when the sign difference · 7bdaecb7
  Chris Lattner authored Aug 06, 2008
  7bdaecb7
- Reinstate this optimization, but without the miscompile. Thanks to Bill for · bf428935
  Nick Lewycky authored Aug 06, 2008
  bf428935
- Just grep for through the LL code instead of the ASM code · 0e966d3e
  Bill Wendling authored Aug 06, 2008
  0e966d3e
- Add default architecture. · bc6786e7
  Bill Wendling authored Aug 05, 2008
  bc6786e7
- Testcase for PR2629. · 3dfa168d
  Bill Wendling authored Aug 05, 2008
  3dfa168d
Aug 05, 2008
- Revert r53282. This was causing a miscompile on Linux. Also, the transformation · ee12a7ae
  Bill Wendling authored Aug 05, 2008
  ee12a7ae
Jul 29, 2008
- Add -unroll-allow-partial command line option that enabled the loop unroller to · 98b5c16e
  Matthijs Kooijman authored Jul 29, 2008
  98b5c16e
- Restructure ArgumentPromotion a bit. Instead of just having a single boolean · fd307045
  Matthijs Kooijman authored Jul 29, 2008
  fd307045
Jul 28, 2008
- Add support for eliminating stores that store the same value that was just loaded. · 3f338974
  Owen Anderson authored Jul 28, 2008
  3f338974
Jul 25, 2008
- Put the LICM of constant GlobalVariables, introduced in r53945, under a · 5f36a32e
  Dan Gohman authored Jul 24, 2008
  5f36a32e
Jul 23, 2008
- "Allow LICM to sink or lift loads from constant memory. Also add a test · 8a8fb908
  Chris Lattner authored Jul 23, 2008
  8a8fb908