Commits · 5dc61c9076d6ad1c9218a10775f939ade8d32fd1 · Raul Torres / llvm-target-spread · GitLab

Mar 29, 2006
- Use unpcklpd for v2f64 splat. · 5dc61c90
  Evan Cheng authored Mar 29, 2006
  5dc61c90
- Floating point logical operation patterns should match bit_convert. Or else · 3cf95747
  Evan Cheng authored Mar 29, 2006
  3cf95747
- Add more SSE intrinsics · 35f1853b
  Evan Cheng authored Mar 29, 2006
  35f1853b
- - More shuffle related bug fixes. · 500ec165
  Evan Cheng authored Mar 29, 2006
  500ec165
- Another entry about shuffles. · 3a1c4e75
  Evan Cheng authored Mar 29, 2006
  3a1c4e75
- - Only use pshufd for v4i32 vector shuffles. · da59b0d2
  Evan Cheng authored Mar 29, 2006
  da59b0d2
- add a note · 7d6f4f14
  Chris Lattner authored Mar 29, 2006
  7d6f4f14
- new testcase · 3cf15bde
  Chris Lattner authored Mar 29, 2006
  3cf15bde
- Bug fixes: handle constantexpr insert/extract element operations · 67271869
  Chris Lattner authored Mar 29, 2006
  67271869
- Added aliases to scalar SSE instructions, e.g. addss, to match x86 intrinsics. · 38b34296
  Evan Cheng authored Mar 28, 2006
  38b34296
- Fixing buggy code. · 8160fd3d
  Evan Cheng authored Mar 28, 2006
  8160fd3d
- Don't sort the names before outputing the intrinsic name table. It causes a · c2c8b58c
  Evan Cheng authored Mar 28, 2006
  c2c8b58c
- When building a VVECTOR_SHUFFLE node from extract_element operations, make · 20e619fb
  Chris Lattner authored Mar 28, 2006
  20e619fb
- Canonicalize VECTOR_SHUFFLE(X, X, Y) -> VECTOR_SHUFFLE(X,undef,Y') · a46dfe80
  Chris Lattner authored Mar 28, 2006
  a46dfe80
Mar 28, 2006
- new testcase · 112cee11
  Chris Lattner authored Mar 28, 2006
  112cee11
- Turn a series of extract_element's feeding a build_vector into a · c9992548
  Chris Lattner authored Mar 28, 2006
  c9992548
- Teach Legalize how to pack VVECTOR_SHUFFLE nodes into VECTOR_SHUFFLE nodes. · f6f94d3b
  Chris Lattner authored Mar 28, 2006
  f6f94d3b
- new node · 8d57da2f
  Chris Lattner authored Mar 28, 2006
  8d57da2f
- Add a new node · 82e86582
  Chris Lattner authored Mar 28, 2006
  82e86582
- Don't crash on X^X if X is a vector. Instead, produce a vector of zeros. · b7163598
  Chris Lattner authored Mar 28, 2006
  b7163598
- Add an assertion · ffec47eb
  Chris Lattner authored Mar 28, 2006
  ffec47eb
- add a note · 66e14108
  Chris Lattner authored Mar 28, 2006
  66e14108
- Refactor address attributes. Add base register to frame info. · dea03488
  Jim Laskey authored Mar 28, 2006
  dea03488
- Expose base register for DwarfWriter. Refactor code accordingly. · d1aa1638
  Jim Laskey authored Mar 28, 2006
  d1aa1638
- More bulletproofing of llvm.dbg.declare. · 67a636c5
  Jim Laskey authored Mar 28, 2006
  67a636c5
- Added missing paren on behalf of Ramana Radhakrishnan. · 457e54ef
  Jim Laskey authored Mar 28, 2006
  457e54ef
- Missed X86::isUNPCKHMask · 21e5476d
  Evan Cheng authored Mar 28, 2006
  21e5476d
- movlps and movlpd should be modeled as two address code. · be2d9a0e
  Evan Cheng authored Mar 28, 2006
  be2d9a0e
- Update · dc57ae07
  Evan Cheng authored Mar 28, 2006
  dc57ae07
- Typo · 4e7374ff
  Evan Cheng authored Mar 28, 2006
  4e7374ff
- * Prefer using operation of matching types. e.g unpcklpd rather than movlhps. · 1a194a52
  Evan Cheng authored Mar 28, 2006
  1a194a52
- Use movhpd is even better than movlhps. · 844a45d5
  Evan Cheng authored Mar 28, 2006
  844a45d5
- Fix a couple typos · af8c373e
  Nate Begeman authored Mar 28, 2006
  af8c373e
- Add a few more altivec intrinsics · 1b392876
  Nate Begeman authored Mar 28, 2006
  1b392876
- Keep Visual Studio informed. · e496f534
  Jeff Cohen authored Mar 28, 2006
  e496f534
- These don't directly map to gcc intrinsics any more. · 9afb95cb
  Chris Lattner authored Mar 28, 2006
  9afb95cb
- Added a couple of entries about movhps and movlhps. · 08b473c6
  Evan Cheng authored Mar 28, 2006
  08b473c6
- All unpack cases are now being handled. · 3765fade
  Evan Cheng authored Mar 28, 2006
  3765fade
- - Clean up / consoladate various shuffle masks. · 2bc32806
  Evan Cheng authored Mar 28, 2006
  2bc32806
- implement a bunch more intrinsics. · 3710fca2
  Chris Lattner authored Mar 28, 2006
  3710fca2