Commits · 8f3b6b8d8a00f6109e1d9c8ef8778d40842d3b11 · Roger Ferrer / llvm-epi-0.8

Apr 04, 2006
- Minor fixes + naming changes. · 8f3b6b8d
  Evan Cheng authored Apr 04, 2006
```
llvm-svn: 27410
```
  8f3b6b8d
- How could this ever have worked? · f00b508e
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27409
```
  f00b508e
- PSHUF* encoding bugs. · 802b35c3
  Evan Cheng authored Apr 04, 2006
```
llvm-svn: 27405
```
  802b35c3
- Make sure to consider alignment of variable sized objects. · e4e7e33c
  Chris Lattner authored Apr 04, 2006
```
This, along with the previous dag combiner fix, fixes
CodeGen/Alpha/2006-04-04-zextload.ll

llvm-svn: 27403
```
  e4e7e33c
- Do not create ZEXTLOAD's unless we are before legalize or the operation is · 4ea52cac
  Chris Lattner authored Apr 04, 2006
```
legal.

llvm-svn: 27402
```
  4ea52cac
- New testcase · d5ed3789
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27401
```
  d5ed3789
- Ask legalize to promote all vector shuffles to be v16i8 instead of having to · 95c7adc7
  Chris Lattner authored Apr 04, 2006
```
handle all 4 PPC vector types.   This simplifies the matching code and allows
us to eliminate a bunch of patterns.  This also adds cases we were missing,
such as CodeGen/PowerPC/vec_splat.ll:splat_h.

llvm-svn: 27400
```
  95c7adc7
- * Add supprot for SCALAR_TO_VECTOR operations where the input needs to be · 6be79823
  Chris Lattner authored Apr 04, 2006
```
  promoted/expanded (e.g. SCALAR_TO_VECTOR from i8/i16 on PPC).
* Add support for targets to request that VECTOR_SHUFFLE nodes be promoted
  to a canonical type, for example, we only want v16i8 shuffles on PPC.
* Move isShuffleLegal out of TLI into Legalize.
* Teach isShuffleLegal to allow shuffles that need to be promoted.

llvm-svn: 27399
```
  6be79823
- Move isShuffleLegal from TLI to Legalize. · 299ecf2a
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27398
```
  299ecf2a
- new testcase · 821fec19
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27397
```
  821fec19
- Signed shr by a constant is not the same as sdiv by 2^k · 6b2c9748
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27395
```
  6b2c9748
- cmpps / cmppd encoding bug · e91e3bd8
  Evan Cheng authored Apr 04, 2006
```
llvm-svn: 27393
```
  e91e3bd8
- Fix the types for these intrinsics. · 3eab63af
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27392
```
  3eab63af
- Constant fold bitconvert(undef) · a9e77d14
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27391
```
  a9e77d14
- Allow targets to have fine grained control over which types various ops get · b2909063
  Chris Lattner authored Apr 04, 2006
```
promoted to, if they desire.

llvm-svn: 27389
```
  b2909063
- Compact some intrinsic definitions. · dd2eb27d
  Evan Cheng authored Apr 04, 2006
```
llvm-svn: 27388
```
  dd2eb27d
- Plug in the byte and short splats · b1e6d845
  Chris Lattner authored Apr 04, 2006
```
llvm-svn: 27387
```
  b1e6d845
- Revert accidentally committed hunks. · 447a7968
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27386
```
  447a7968
- Make sure to mark unsupported SCALAR_TO_VECTOR operations as expand. · 533aed9a
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27385
```
  533aed9a
- Some SSE1 intrinsics: min, max, sqrt, etc. · 0ef83c83
  Evan Cheng authored Apr 03, 2006
```
llvm-svn: 27384
```
  0ef83c83
- revert previous patch · bf0016f2
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27383
```
  bf0016f2
- Use movlpd to: store lower f64 extracted from v2f64. · b64827e6
  Evan Cheng authored Apr 03, 2006
```
Use movhpd to: store upper f64 extracted from v2f64.

llvm-svn: 27382
```
  b64827e6
- Force use of a frame-pointer if there is anything on the stack that is aligned · 54007275
  Chris Lattner authored Apr 03, 2006
```
more than the OS keeps the stack aligned.

llvm-svn: 27381
```
  54007275
Apr 03, 2006
- The stack alignment is now computed dynamically, just verify it is correct. · b710a81e
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27380
```
  b710a81e
- Remove unused method · 6bc4b9c7
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27379
```
  6bc4b9c7
- Keep track of max stack alignment as objects are added. Remove an obsolete method. · 6dc98705
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27378
```
  6dc98705
- - More efficient extract_vector_elt with shuffle and movss, movsd, movd, etc. · ebf1006d
  Evan Cheng authored Apr 03, 2006
```
- Some bug fixes and naming inconsistency fixes.

llvm-svn: 27377
```
  ebf1006d
- Align vectors to the size in bytes, not bits. · 78c788b4
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27376
```
  78c788b4
- Add a missing check, this fixes UnitTests/Vector/sumarray.c · e1e3adf8
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27375
```
  e1e3adf8
- Add a missing check, which broke a bunch of vector tests. · 04c00fc8
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27374
```
  04c00fc8
- shrinkify intrinsics more by using some local classes · d97f0389
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27373
```
  d97f0389
- Add the full set of min/max instructions · 9ccd61c8
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27372
```
  9ccd61c8
- Add some classes to make it easier to define intrinsics. Add min/max intrinsics. · 36a519b0
  Chris Lattner authored Apr 03, 2006
```
llvm-svn: 27371
```
  36a519b0
- support x * (c1 + c2) where c1 and c2 are pow2s. special case for c2 == 4 · df7abf8b
  Andrew Lenharth authored Apr 03, 2006
```
llvm-svn: 27370
```
  df7abf8b
- test powers of 2 · 2636d2ac
  Andrew Lenharth authored Apr 03, 2006
```
llvm-svn: 27369
```
  2636d2ac
- mul by const conversion sequences. more coming soon · 4e2c073a
  Andrew Lenharth authored Apr 03, 2006
```
llvm-svn: 27368
```
  4e2c073a
- back this out · 94f012f6
  Andrew Lenharth authored Apr 03, 2006
```
llvm-svn: 27367
```
  94f012f6
- test some more mul by constant removal · 0288ba76
  Andrew Lenharth authored Apr 03, 2006
```
llvm-svn: 27366
```
  0288ba76
Apr 02, 2006
- Make sure mul by constant 5 is turned into a s4addq · 04a84295
  Andrew Lenharth authored Apr 02, 2006
```
llvm-svn: 27365
```
  04a84295
- This should be a win of every arch · 015eaf5f
  Andrew Lenharth authored Apr 02, 2006
```
llvm-svn: 27364
```
  015eaf5f