Commits · 74cf9ff761bcb8755af8c08d9ab0ec62ba33195d · Roger Ferrer / llvm-epi-0.8

Apr 12, 2006
- Rename get_VSPLI_elt -> get_VSPLTI_elt · 74cf9ff7
  Chris Lattner authored Apr 12, 2006
```
Canonicalize BUILD_VECTOR's that match VSPLTI's into a single type for each
form, eliminating a bunch of Pat patterns in the .td file and allowing us to
CSE stuff more aggressively.  This implements
PowerPC/buildvec_canonicalize.ll:VSPLTI

llvm-svn: 27614
```
  74cf9ff7
- Promote v4i32, v8i16, v16i8 load to v2i64 load. · e2157c6e
  Evan Cheng authored Apr 12, 2006
```
llvm-svn: 27612
```
  e2157c6e
- Ensure that zero vectors are always v4i32, which forces them to CSE with · e318a757
  Chris Lattner authored Apr 12, 2006
```
each other.  This implements CodeGen/PowerPC/vxor-canonicalize.ll

llvm-svn: 27609
```
  e318a757
- Various SSE2 conversion intrinsics · 29be057d
  Evan Cheng authored Apr 12, 2006
```
llvm-svn: 27603
```
  29be057d
- Added __builtin_ia32_storelv4si, __builtin_ia32_movqv4si, · 70c74a3c
  Evan Cheng authored Apr 11, 2006
```
__builtin_ia32_loadlv4si, __builtin_ia32_loaddqu, __builtin_ia32_storedqu.

llvm-svn: 27599
```
  70c74a3c
Apr 11, 2006
- Fix SingleSource/UnitTests/Vector/sumarray-dbl · f19bcd51
  Nate Begeman authored Apr 11, 2006
```
llvm-svn: 27594
```
  f19bcd51
- Fix PR727, correctly handling large stack aligments on ppc · 1bb13209
  Nate Begeman authored Apr 11, 2006
```
llvm-svn: 27593
```
  1bb13209
- we have a shuffle instr, add an example. · aaa04230
  Chris Lattner authored Apr 11, 2006
```
llvm-svn: 27592
```
  aaa04230
- gcc lower SSE prefetch into generic prefetch intrinsic. Need to add support · 6b60357f
  Evan Cheng authored Apr 11, 2006
```
later.

llvm-svn: 27591
```
  6b60357f
- Misc. intrinsics. · 6ea715af
  Evan Cheng authored Apr 11, 2006
```
llvm-svn: 27590
```
  6ea715af
- Suppress debug label when not debug. · 02b3b72b
  Jim Laskey authored Apr 11, 2006
```
llvm-svn: 27588
```
  02b3b72b
- movnt* and maskmovdqu intrinsics · 09a95627
  Evan Cheng authored Apr 11, 2006
```
llvm-svn: 27587
```
  09a95627
- Vector function results go into V2 according to GCC. The darwin ABI doc · e4db08a2
  Chris Lattner authored Apr 11, 2006
```
doesn't say where they go :-/

llvm-svn: 27579
```
  e4db08a2
- Move some return-handling code from lowerarguments to the ISD::RET handling stuff. · 92533cfb
  Chris Lattner authored Apr 11, 2006
```
No functionality change.

llvm-svn: 27577
```
  92533cfb
- Added support for _mm_move_ss and _mm_move_sd. · 12ba3e23
  Evan Cheng authored Apr 11, 2006
```
llvm-svn: 27575
```
  12ba3e23
- Use existing information. · dca2655d
  Jim Laskey authored Apr 10, 2006
```
llvm-svn: 27574
```
  dca2655d
- Remove some bogus patterns; clean up. · f8ac0228
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27569
```
  f8ac0228
Apr 10, 2006
- add a note · d99f57c1
  Chris Lattner authored Apr 10, 2006
```
llvm-svn: 27567
```
  d99f57c1
- Remove an entry that is now done. · 051de9a8
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27565
```
  051de9a8
- Added some missing shuffle patterns. · 76112c3c
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27564
```
  76112c3c
- Correct an entry · 664fcba5
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27563
```
  664fcba5
- movups / movupd · 395fa3d2
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27562
```
  395fa3d2
- Conditional move of vector types. · 617a6a81
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27556
```
  617a6a81
- New entries · 014849e1
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27555
```
  014849e1
- Use movaps to do VR128 reg-to-reg copies for now. It's shorter and available for SSE1. · c9ed8e4c
  Evan Cheng authored Apr 10, 2006
```
llvm-svn: 27554
```
  c9ed8e4c
Apr 09, 2006
- properly mark vector selects as expanded to select_cc · 3a68f3c3
  Chris Lattner authored Apr 08, 2006
```
llvm-svn: 27544
```
  3a68f3c3
- Add VRRC select support · 0a3d1bbc
  Chris Lattner authored Apr 08, 2006
```
llvm-svn: 27543
```
  0a3d1bbc
Apr 08, 2006
- Disable switch lowering for targets based on the selection dag isel, · 3f9c1790
  Nate Begeman authored Apr 08, 2006
```
letting the code generator handle them directly.

llvm-svn: 27539
```
  3f9c1790
- Implement PowerPC/CodeGen/vec_splat.ll:spltish to use vsplish instead of a · d9e80f45
  Chris Lattner authored Apr 08, 2006
```
constant pool load.

llvm-svn: 27538
```
  d9e80f45
- Change the interface to the predicate that determines if vsplti* can be used. · d71a1f94
  Chris Lattner authored Apr 08, 2006
```
No functionality changes.

llvm-svn: 27536
```
  d71a1f94
- Initialize SDOperand values because the gcc 4.0.2 compiler complains about · cf905223
  Reid Spencer authored Apr 08, 2006
```
them.

llvm-svn: 27534
```
  cf905223
- ldmxcsr and stmxcsr. · 0df9c9f5
  Evan Cheng authored Apr 08, 2006
```
llvm-svn: 27506
```
  0df9c9f5
Apr 07, 2006
- Code clean up. · ac847268
  Evan Cheng authored Apr 07, 2006
```
llvm-svn: 27501
```
  ac847268
- Added patterns for MOVHPSmr and MOVLPSmr. · aa18a525
  Evan Cheng authored Apr 07, 2006
```
llvm-svn: 27497
```
  aa18a525
- Keep track of an Mac OS X / x86 ABI bug. · 748e573c
  Evan Cheng authored Apr 07, 2006
```
llvm-svn: 27496
```
  748e573c
- Make sure that debug labels are defined within the same section and after the · c0d6518f
  Jim Laskey authored Apr 07, 2006
```
entry point of a function.

llvm-svn: 27494
```
  c0d6518f
- Foundation for call frame information. · 2d7298c3
  Jim Laskey authored Apr 07, 2006
```
llvm-svn: 27491
```
  2d7298c3
- A MOVPS2SSmr, i.e. _mm_store_ss, encoding bug. · d8e1a01b
  Evan Cheng authored Apr 06, 2006
```
Also MOVPDI2DIrr.

llvm-svn: 27476
```
  d8e1a01b
- - movlp{s|d} and movhp{s|d} support. · c995b45f
  Evan Cheng authored Apr 06, 2006
```
- Normalize shuffle nodes so result vector lower half elements come from the
  first vector, the rest come from the second vector. (Except for the
  exceptions :-).
- Other minor fixes.

llvm-svn: 27474
```
  c995b45f
- New entries. · acf8b3c8
  Evan Cheng authored Apr 06, 2006
```
llvm-svn: 27473
```
  acf8b3c8