Commits · 63fab495bb318a16624f0acd3e8ca4e6fc7b97ae · Roger Ferrer / llvm-epi · GitLab

Sep 02, 2005
- Test that converting from double to int64 results in one libcall, not one · 63fab495
  Chris Lattner authored Sep 02, 2005
  63fab495
- turn on dag isel by default · 06e237f2
  Chris Lattner authored Sep 02, 2005
  06e237f2
- Make sure to auto-cse nullary ops · b95b280b
  Chris Lattner authored Sep 02, 2005
  b95b280b
- add a map for nullary ops · 5ba7b615
  Chris Lattner authored Sep 02, 2005
  5ba7b615
- Add help support for -mcpu and -mattr. · 27d628df
  Jim Laskey authored Sep 02, 2005
  27d628df
- Fix some buggy logic where we would try to remove nodes with two operands · 1e89e36d
  Chris Lattner authored Sep 02, 2005
  1e89e36d
- Pull out Lowering in preperation for multiple ISels. Oh, and get rid of some stuff · 9690a4f3
  Andrew Lenharth authored Sep 02, 2005
  9690a4f3
- Don't create zero sized stack objects even for array allocas with a zero · b0b4ec56
  Chris Lattner authored Sep 02, 2005
  b0b4ec56
- Decouple fsqrt from gpul optimizations, implementing fsqrt.ll. · aa3b1fcc
  Chris Lattner authored Sep 02, 2005
  aa3b1fcc
- new testcase to ensure fsqrt is generated for correct subtargets only, and · ffb99034
  Chris Lattner authored Sep 02, 2005
  ffb99034
- Move a bunch of non-deprecated methods above the "deprecated line" · 38ad3f69
  Chris Lattner authored Sep 02, 2005
  38ad3f69
- Fix the release build, noticed by Eric van Riet Paap · b6cde17d
  Chris Lattner authored Sep 02, 2005
  b6cde17d
- Fix a problem that Dan Berlin noticed, where reassociation would not succeed · b5e381a8
  Chris Lattner authored Sep 02, 2005
  b5e381a8
- Avoid creating garbage instructions, just move the old add instruction · 9fe263aa
  Chris Lattner authored Sep 02, 2005
  9fe263aa
- new testcase for recent bugfix · b944d5a4
  Chris Lattner authored Sep 02, 2005
  b944d5a4
- add some assertions and fix problems where reassociate could access the · d1325da0
  Chris Lattner authored Sep 02, 2005
  d1325da0
- Fix VC++ build errors · a6dde996
  Jeff Cohen authored Sep 02, 2005
  a6dde996
- Restore this patch now that the latent bug has been fixed · 763a3a0f
  Chris Lattner authored Sep 02, 2005
  763a3a0f
- Make sure to legalize assert[zs]ext's operand correctly · d9af1aab
  Chris Lattner authored Sep 02, 2005
  d9af1aab
- Revert the previous patch which causes a mysterious regression in toast. · 06d440f2
  Chris Lattner authored Sep 02, 2005
  06d440f2
- Teach live intervals to not crash on dead livein regs · 7138f914
  Chris Lattner authored Sep 02, 2005
  7138f914
- For values that are live across basic blocks and need promotion, use ANY_EXTEND · a66403db
  Chris Lattner authored Sep 02, 2005
  a66403db
- legalize ANY_EXTEND appropriately · 7753f175
  Chris Lattner authored Sep 02, 2005
  7753f175
- Add support for ANY_EXTEND and add a few minor folds for it · 8c393c21
  Chris Lattner authored Sep 02, 2005
  8c393c21
- Handle any_extend like zext · 210975cf
  Chris Lattner authored Sep 02, 2005
  210975cf
- Handle ANY_EXTEND like ZERO_EXTEND. Simplify the extend/truncate code on · 2493f0e5
  Chris Lattner authored Sep 02, 2005
  2493f0e5
- Add a new ANY_EXTEND node, which operates like an extension but has undefined · 969af57d
  Chris Lattner authored Sep 02, 2005
  969af57d
- This should permit NOT and's, not even dead ones. · 97ac77fd
  Chris Lattner authored Sep 02, 2005
  97ac77fd
- Implement small-arguments.ll:test3 by teaching the DAG optimizer that · 9ee867b9
  Chris Lattner authored Sep 01, 2005
  9ee867b9
- Another case that shouldn't have sign extends: functions returning short · 907123b1
  Chris Lattner authored Sep 01, 2005
  907123b1
- Fix some code in the current node combining code, spotted when it was moved · d78d9754
  Nate Begeman authored Sep 01, 2005
  d78d9754
- Implement first round of feedback from chris (there's still a couple things · 2504fe26
  Nate Begeman authored Sep 01, 2005
  2504fe26
- Align functions to 16-byte boundaries, to eliminate noise in performance... · 68d15fdf
  Chris Lattner authored Sep 01, 2005
  68d15fdf
Sep 01, 2005
- Local labels on darwin apparently start with just 'L', not .L like other · e40a3ccd
  Chris Lattner authored Sep 01, 2005
  e40a3ccd
- · 19058c39
  Jim Laskey authored Sep 01, 2005
  19058c39
- This new class provides support for platform specific "features". The intent · 3fee6a51
  Jim Laskey authored Sep 01, 2005
  3fee6a51
- Implement dynamic allocas correctly. In particular, because we were copying · a305d28c
  Chris Lattner authored Sep 01, 2005
  a305d28c
- Fix a bug where we were useing HA to get the high part, which seems like it · 293b3a68
  Chris Lattner authored Sep 01, 2005
  293b3a68
- Do not select the operands being passed into SelectCC. IT does this itself · 34182aff
  Chris Lattner authored Sep 01, 2005
  34182aff
- It is NDEBUG not _NDEBUG · 975f5c9f
  Chris Lattner authored Sep 01, 2005
  975f5c9f