Commits · c404b5748cf74ca34eca33e82f8a80199f082f59 · Roger Ferrer / llvm-epi-0.8

Feb 01, 2006
- Remove another entry. · c404b574
  Evan Cheng authored Feb 01, 2006
```
llvm-svn: 25871
```
  c404b574
- Another regression from the pattern isel · b0a76b09
  Chris Lattner authored Feb 01, 2006
```
llvm-svn: 25867
```
  b0a76b09
- Return's chain should be matching either the chain produced by the · a24617f5
  Evan Cheng authored Feb 01, 2006
```
value or the chain going into the load.

llvm-svn: 25863
```
  a24617f5
- When folding a load into a return of SSE value, check the chain to · e1ce4d71
  Evan Cheng authored Feb 01, 2006
```
ensure the memory location has not been clobbered.

llvm-svn: 25861
```
  e1ce4d71
- Remove an item. It's done. · bc1fcd07
  Evan Cheng authored Feb 01, 2006
```
llvm-svn: 25860
```
  bc1fcd07
- Be smarter about whether to store the SSE return value in memory. If · 5659ca8f
  Evan Cheng authored Jan 31, 2006
```
it is already available in memory, do a fld directly from there.

llvm-svn: 25859
```
  5659ca8f
Jan 31, 2006
- turning these into 'adds' would require extra copies · 64387c3e
  Chris Lattner authored Jan 31, 2006
```
llvm-svn: 25858
```
  64387c3e
- - Allow XMM load (for scalar use) to be folded into ANDP* and XORP*. · 72d5c256
  Evan Cheng authored Jan 31, 2006
```
- Use XORP* to implement fneg.

llvm-svn: 25857
```
  72d5c256
- Remove entries on fabs and fneg. These are done. · a91eb485
  Evan Cheng authored Jan 31, 2006
```
llvm-svn: 25856
```
  a91eb485
- * Fix 80-column violations · c642aa5e
  Chris Lattner authored Jan 31, 2006
```
* Rename hasSSE -> hasSSE1 to avoid my continual confusion with 'has any SSE'.
* Add inline asm constraint specification.

llvm-svn: 25854
```
  c642aa5e
- Added custom lowering of fabs · 2dd217b8
  Evan Cheng authored Jan 31, 2006
```
llvm-svn: 25831
```
  2dd217b8
- Another high-prio selection performance bug · d916e78b
  Chris Lattner authored Jan 31, 2006
```
llvm-svn: 25828
```
  d916e78b
- more mumbling · 2b70a6f8
  Chris Lattner authored Jan 31, 2006
```
llvm-svn: 25826
```
  2b70a6f8
- add some notes · b521361f
  Chris Lattner authored Jan 31, 2006
```
llvm-svn: 25825
```
  b521361f
- Don't generate complex sequence for SETOLE, SETOLT, SETULT, and SETUGT. Flip · 45df7f84
  Evan Cheng authored Jan 30, 2006
```
the order of the compare operands and generate SETOGT, SETOGE, SETUGE, and
SETULE instead.

llvm-svn: 25824
```
  45df7f84
Jan 30, 2006
- i64 -> f32, f32 -> i64 and some clean up. · 08390f6a
  Evan Cheng authored Jan 30, 2006
```
llvm-svn: 25818
```
  08390f6a
- Always use FP stack instructions to perform i64 to f64 as well as f64 to i64 · 5b97fcf0
  Evan Cheng authored Jan 30, 2006
```
conversions. SSE does not have instructions to handle these tasks.

llvm-svn: 25817
```
  5b97fcf0
- Move MaskedValueIsZero from the DAGCombiner to the TargetLowering... · f0b24d2d
  Chris Lattner authored Jan 30, 2006
```
Move MaskedValueIsZero from the DAGCombiner to the TargetLowering interface,making isMaskedValueZeroForTargetNode simpler, and useable from other partsof the compiler.

llvm-svn: 25803
```
  f0b24d2d
- adjust prototype · c6fa0282
  Chris Lattner authored Jan 30, 2006
```
llvm-svn: 25798
```
  c6fa0282
Jan 29, 2006
- add another note · 3c6a9506
  Chris Lattner authored Jan 29, 2006
```
llvm-svn: 25789
```
  3c6a9506
- add some performance notes from looking at sgefa · dabee1f6
  Chris Lattner authored Jan 29, 2006
```
llvm-svn: 25788
```
  dabee1f6
- add a high-priority SSE issue from sgefa · 7c7cbde0
  Chris Lattner authored Jan 29, 2006
```
llvm-svn: 25787
```
  7c7cbde0
- add a missed optimization · 5a7a22c9
  Chris Lattner authored Jan 29, 2006
```
llvm-svn: 25786
```
  5a7a22c9
- Add a note about lowering llvm.memset, llvm.memcpy, and llvm.memmove to a · 0c05a2c9
  Reid Spencer authored Jan 29, 2006
```
few stores under certain conditions.

llvm-svn: 25777
```
  0c05a2c9
- remove now-dead code, the legalizer takes care of this for us · 35d20a4c
  Chris Lattner authored Jan 29, 2006
```
llvm-svn: 25776
```
  35d20a4c
- The FP stack doesn't support UNDEF, ask the legalizer to legalize it · 132177e1
  Chris Lattner authored Jan 29, 2006
```
instead of lying and saying we have it.

llvm-svn: 25775
```
  132177e1
- Targets all now request ConstantFP to be legalized into TargetConstantFP. · 61c9a8e9
  Chris Lattner authored Jan 29, 2006
```
'fpimm' in .td files is now TargetConstantFP.

llvm-svn: 25771
```
  61c9a8e9
- Fix typo. · 4ab39e43
  Jeff Cohen authored Jan 29, 2006
```
llvm-svn: 25760
```
  4ab39e43
Jan 28, 2006
- Flesh out AMD family/models. · 8643ea67
  Jeff Cohen authored Jan 28, 2006
```
llvm-svn: 25755
```
  8643ea67
- Correctly determine CPU vendor. · 58ca0be9
  Jeff Cohen authored Jan 28, 2006
```
llvm-svn: 25754
```
  58ca0be9
- Use union instead of reinterpret_cast. · 71287085
  Jeff Cohen authored Jan 28, 2006
```
llvm-svn: 25751
```
  71287085
- Fix recognition of Intel CPUs. · b5de47cd
  Jeff Cohen authored Jan 28, 2006
```
llvm-svn: 25750
```
  b5de47cd
- Is64Bit reflects the capability of the chip, not an aspect of the target os · b3ab2d3a
  Chris Lattner authored Jan 28, 2006
```
llvm-svn: 25749
```
  b3ab2d3a
- Fix a bunch of JIT failures with the new isel · be08957d
  Chris Lattner authored Jan 28, 2006
```
llvm-svn: 25748
```
  be08957d
- Improve X86 subtarget support for Windows and AMD. · e128d5f7
  Jeff Cohen authored Jan 28, 2006
```
llvm-svn: 25747
```
  e128d5f7
- silence a warning · ccd2a20c
  Chris Lattner authored Jan 28, 2006
```
llvm-svn: 25745
```
  ccd2a20c
- make this work on non-native hosts · dc8bbb65
  Chris Lattner authored Jan 28, 2006
```
llvm-svn: 25734
```
  dc8bbb65
Jan 27, 2006
- A bit of wisdom from Chris on the last entry. · 18243826
  Evan Cheng authored Jan 27, 2006
```
llvm-svn: 25715
```
  18243826
- AT&T assembly convention: registers are in lower case. · 63045d22
  Evan Cheng authored Jan 27, 2006
```
llvm-svn: 25714
```
  63045d22
- initialize all instance vars · dbfc2999
  Chris Lattner authored Jan 27, 2006
```
llvm-svn: 25711
```
  dbfc2999