Commits · f7bde9a1927a3a464e4159863b6f73b42d6c82c6 · Roger Ferrer / llvm-epi · GitLab

Jan 24, 2011
- merge all the "crash tests" into crash.ll · 2bcec129
  Chris Lattner authored Jan 24, 2011
  2bcec129
- fix PR9017, a bug where we'd assert when promoting in unreachable · b4017769
  Chris Lattner authored Jan 24, 2011
  b4017769
- enhance SRoA to promote allocas that are used by PHI nodes. This often · d83e7b0f
  Chris Lattner authored Jan 24, 2011
  d83e7b0f
Jan 23, 2011
- Enhance SRoA to promote allocas that are used by selects in some · a960725d
  Chris Lattner authored Jan 23, 2011
  a960725d
- Enhance SRoA to be more aggressive about scalarization of aggregate allocas · 9491dee2
  Chris Lattner authored Jan 23, 2011
  9491dee2
- remove an old hack that avoided creating MMX datatypes. The · a587ab7b
  Chris Lattner authored Jan 23, 2011
  a587ab7b
Jan 21, 2011
- Actually check memcpy lengths, instead of just commenting about · 19e30d5a
  Dan Gohman authored Jan 21, 2011
  19e30d5a
- Just because we have determined that an (fcmp | fcmp) is true for A < B, · a834200d
  Owen Anderson authored Jan 21, 2011
  a834200d
- fix PR9013, an infinite loop in instcombine. · b5e15d19
  Chris Lattner authored Jan 21, 2011
  b5e15d19
- Don't try to pull vector bitcasts that change the number of elements through · 6a083cf8
  Nick Lewycky authored Jan 21, 2011
  6a083cf8
- Add a constant folding of casts from zero to zero. Fixes PR9011! · 39b12c05
  Nick Lewycky authored Jan 21, 2011
  39b12c05
Jan 20, 2011
- At -O123 the early-cse pass is run before instcombine has run. According to my · 8fb2c382
  Duncan Sands authored Jan 20, 2011
  8fb2c382
Jan 19, 2011
- Add unnamed_addr when we can show that address of a global is not used. · fc355bc0
  Rafael Espindola authored Jan 19, 2011
  fc355bc0
Jan 18, 2011
- For completeness, generalize the (X + Y) - Y -> X transform and add X - (X + 1) -> -1. · 99589d07
  Duncan Sands authored Jan 18, 2011
  99589d07
- Simplify (X<<1)-X into X. According to my auto-simplier this is the most common missed · 9b8e2bd8
  Duncan Sands authored Jan 18, 2011
  9b8e2bd8
Jan 16, 2011
- Test for lazy value info's ability to prove the absense of NULLs in pointers. · 872a453a
  Nick Lewycky authored Jan 16, 2011
  872a453a
- Teach DAE to look for functions whose arguments are unused, and change all... · d3db8334
  Anders Carlsson authored Jan 16, 2011
  d3db8334
- Don't merge two constants if we care about the address of both. · 751677a0
  Rafael Espindola authored Jan 16, 2011
  751677a0
- Reduce and merge testcases. · ec3b10fc
  Owen Anderson authored Jan 16, 2011
  ec3b10fc
- fix PR8932, a case where arg promotion could infinitely promote. · e5f8de86
  Chris Lattner authored Jan 16, 2011
  e5f8de86
- if an alloca is only ever accessed as a unit, and is accessed with load/store instructions, · 6fab2e94
  Chris Lattner authored Jan 16, 2011
  6fab2e94
- enhance FoldOpIntoPhi in instcombine to try harder when a phi has · d55581de
  Chris Lattner authored Jan 16, 2011
  d55581de
- Improve the safety of my globalopt enhancement by ensuring that the bitcast · 4e54efd6
  Owen Anderson authored Jan 16, 2011
  4e54efd6
- fix PR8983, a broken assertion. · 08f43456
  Chris Lattner authored Jan 16, 2011
  08f43456
- remove the partial specialization pass. It is unmaintained and has bugs. · 1e209b87
  Chris Lattner authored Jan 16, 2011
  1e209b87
Jan 15, 2011
- Make constmerge a two-pass algorithm so that it won't miss merging · 0296a481
  Nick Lewycky authored Jan 15, 2011
  0296a481
- temporarily revert r123526. While working on a follow-on patch I · af263907
  Chris Lattner authored Jan 15, 2011
  af263907
- fix rdar://8785296 - -fcatch-undefined-behavior generates inefficient code · 8df83c4a
  Chris Lattner authored Jan 15, 2011
  8df83c4a
- Now that instruction optzns can update the iterator as they go, we can · 1b93be50
  Chris Lattner authored Jan 15, 2011
  1b93be50
- implement an instcombine xform that canonicalizes casts outside of and-with-constant operations. · 9c10d587
  Chris Lattner authored Jan 15, 2011
  9c10d587
Jan 14, 2011
- Turn X-(X-Y) into Y. According to my auto-simplifier this is the most common · d6f1a958
  Duncan Sands authored Jan 14, 2011
  d6f1a958
- Factorize common code out of the InstructionSimplify shift logic. Add in · 571fd9a6
  Duncan Sands authored Jan 14, 2011
  571fd9a6
- Rename this test. · c3eb0f4b
  Duncan Sands authored Jan 14, 2011
  c3eb0f4b
- relax testcase a bit. · 5e0fef85
  Chris Lattner authored Jan 14, 2011
  5e0fef85
- Move some shift transforms out of instcombine and into InstructionSimplify. · 7f60dc1e
  Duncan Sands authored Jan 14, 2011
  7f60dc1e
Jan 13, 2011
- Extend SROA to handle arrays accessed as homogeneous structs and vice versa. · 08713d3c
  Bob Wilson authored Jan 13, 2011
  08713d3c
- Make SROA more aggressive with allocas containing padding. · 12eec40c
  Bob Wilson authored Jan 13, 2011
  12eec40c
- The most common simplification missed by instsimplify in unoptimized bitcode · 8d25a7c3
  Duncan Sands authored Jan 13, 2011
  8d25a7c3
Jan 12, 2011
- revert 123144, reenabling the rest of memset formation. · dd5f60b7
  Chris Lattner authored Jan 12, 2011
  dd5f60b7
- revert r123146 which disabled code that wasn't the root cause · 654098f4
  Chris Lattner authored Jan 12, 2011
  654098f4