Commits · fa57a0add935b26c7362af4c90d30588e0e19549 · Roger Ferrer / llvm-epi-0.8

Feb 21, 2006
- Added SSE2 128-bit integer packed types: V16I8, V8I16, V4I32, and V2I64. · fa57a0ad
  Evan Cheng authored Feb 21, 2006
```
Added generic vector types: VR64 and VR128.

llvm-svn: 26295
```
  fa57a0ad
Feb 20, 2006
- Added x86 integer vector types: 64-bit packed byte integer (v16i8), 64-bit · 43070b75
  Evan Cheng authored Feb 20, 2006
```
packed word integer (v8i16), and 64-bit packed doubleword integer (v2i32).

llvm-svn: 26294
```
  43070b75
- Some updates · 4547400a
  Evan Cheng authored Feb 20, 2006
```
llvm-svn: 26292
```
  4547400a
- Fix a problem Nate and Duraid reported where simplifying nodes can cause · 301f45cf
  Chris Lattner authored Feb 20, 2006
```
them to get ressurected, in which case, deleting the undead nodes is
unfriendly.

llvm-svn: 26291
```
  301f45cf
- Fix a problem on itanium with memset. The value to set has been promoted to · 486d1bc5
  Chris Lattner authored Feb 20, 2006
```
i64 before this code, so zero_ext doesn't work.

llvm-svn: 26290
```
  486d1bc5
Feb 18, 2006
- improved support for branch folding, still not enabled. · c2e3a7a4
  Chris Lattner authored Feb 18, 2006
```
llvm-svn: 26289
```
  c2e3a7a4
- If SSE3 is available, promote FP_TO_UINT i32 to FP_TO_SINT i64 to take · d13778eb
  Evan Cheng authored Feb 18, 2006
```
advantage of fisttpll.

llvm-svn: 26288
```
  d13778eb
- Fix bugs identified by VC++. · 0add83e9
  Jeff Cohen authored Feb 18, 2006
```
llvm-svn: 26287
```
  0add83e9
- Add a fold for add that exchanges it with a constant shift if possible, so · 983ca897
  Nate Begeman authored Feb 18, 2006
```
that the shift may be more easily folded into other operations.

llvm-svn: 26286
```
  983ca897
- Implement deletion of dead blocks, currently disabled. · 19fa8ac9
  Chris Lattner authored Feb 18, 2006
```
llvm-svn: 26285
```
  19fa8ac9
- Add checks to make sure we don't create bogus extend nodes, and fix a bug · abac6160
  Nate Begeman authored Feb 18, 2006
```
where we were doing exactly that which was causing failures on x86 and
alpha.

llvm-svn: 26284
```
  abac6160
- Added fisttp for fp to int conversion. · 70af6207
  Evan Cheng authored Feb 18, 2006
```
llvm-svn: 26283
```
  70af6207
- Disable PIC for JIT. · 06c2e6d1
  Evan Cheng authored Feb 18, 2006
```
llvm-svn: 26281
```
  06c2e6d1
- a previous patch completely disabled trivial unswitching, this fixees it. · cb853de5
  Chris Lattner authored Feb 18, 2006
```
Thanks to nate for pointing this out :)

llvm-svn: 26280
```
  cb853de5
- initial trivial support for folding branches that have now-constant destinations. · 29f771ba
  Chris Lattner authored Feb 18, 2006
```
llvm-svn: 26279
```
  29f771ba
- Jit does not support PIC yet. · 5caed8a2
  Evan Cheng authored Feb 18, 2006
```
llvm-svn: 26278
```
  5caed8a2
- When unswitching a loop, make sure to update loop info with exit blocks in · 8e44ff50
  Chris Lattner authored Feb 18, 2006
```
the right loop.

llvm-svn: 26277
```
  8e44ff50
- Fix Transforms/SimplifyCFG/2006-02-17-InfiniteUnroll.ll · d9566518
  Chris Lattner authored Feb 18, 2006
```
llvm-svn: 26275
```
  d9566518
- x86 / Darwin PIC support. · 5588de94
  Evan Cheng authored Feb 18, 2006
```
llvm-svn: 26273
```
  5588de94
- Moved PICEnabled to include/llvm/Target/TargetOptions.h · 5f99760a
  Evan Cheng authored Feb 18, 2006
```
llvm-svn: 26272
```
  5f99760a
Feb 17, 2006
- Fix a tricky issue in the SimplifyDemandedBits code where CombineTo wasn't · 375e1a71
  Chris Lattner authored Feb 17, 2006
```
exactly the API we wanted to call into.  This fixes the crash on crafty last
night.

llvm-svn: 26269
```
  375e1a71
- Clean up DemandedBitsAreZero interface · fb5dbadf
  Nate Begeman authored Feb 17, 2006
```
Make more use of the new mask helpers in valuetypes.h
Combine (sra (srl x, c1), c1) -> sext_inreg if legal

llvm-svn: 26263
```
  fb5dbadf
- Don't expand sdiv by power of two before legalize, since it will likely · 57b35675
  Nate Begeman authored Feb 17, 2006
```
generate illegal nodes.

llvm-svn: 26261
```
  57b35675
- unbreak the build · 07a2677e
  Chris Lattner authored Feb 17, 2006
```
llvm-svn: 26260
```
  07a2677e
- Unbreak x86 be · 593bea73
  Evan Cheng authored Feb 17, 2006
```
llvm-svn: 26259
```
  593bea73
- Fix loops where the header has an exit, fixing a loop-unswitch crash on crafty · baddba41
  Chris Lattner authored Feb 17, 2006
```
llvm-svn: 26258
```
  baddba41
- kill ADD_PARTS & SUB_PARTS and replace them with fancy new ADDC, ADDE, SUBC · 5965bd19
  Nate Begeman authored Feb 17, 2006
```
and SUBE nodes that actually expose what's going on and allow for
significant simplifications in the targets.

llvm-svn: 26255
```
  5965bd19
- Fix another miscompilation exposed by lencode, where we lowered i64->f32 · 9ec392b2
  Chris Lattner authored Feb 17, 2006
```
conversions to __floatdidf instead of __floatdisf on targets that support
f32 but not i64 (e.g. sparc).

llvm-svn: 26254
```
  9ec392b2
- add note about div by power of 2 · 67c21b6c
  Chris Lattner authored Feb 17, 2006
```
llvm-svn: 26253
```
  67c21b6c
- Fix bug noticed by VC++. · 0d62ebd1
  Jeff Cohen authored Feb 17, 2006
```
llvm-svn: 26252
```
  0d62ebd1
- Whoops, didn't mean to check this in yet. · 3920ce4d
  Nate Begeman authored Feb 17, 2006
```
llvm-svn: 26250
```
  3920ce4d
- Add a missing and useful pat frag · 4a0dc0c8
  Nate Begeman authored Feb 17, 2006
```
llvm-svn: 26249
```
  4a0dc0c8
- start of some new simplification code, not thoroughly tested, use at your own · 6fd13623
  Chris Lattner authored Feb 17, 2006
```
risk :)

llvm-svn: 26248
```
  6fd13623
- Remind ourselves to revisit the "pxor vs. xorps/xorpd to clear XMM registers" · b590d3a7
  Evan Cheng authored Feb 17, 2006
```
issue. Need to do more experiments.

llvm-svn: 26247
```
  b590d3a7
- Kill the x86 pattern isel. boom. · 7e5496d5
  Nate Begeman authored Feb 17, 2006
```
llvm-svn: 26246
```
  7e5496d5
- Remove the entry about using movapd for SSE reg-reg moves. · db1dbbe8
  Evan Cheng authored Feb 17, 2006
```
llvm-svn: 26245
```
  db1dbbe8
- pxor (for FLD0SS) encoding was missing the OpSize prefix. · eb7b3380
  Evan Cheng authored Feb 16, 2006
```
llvm-svn: 26244
```
  eb7b3380
- Remove the skeleton target, it doesn't produce useful code and there are · 936cc9fe
  Chris Lattner authored Feb 16, 2006
```
other small targets that do that can be learned from.  They also have
the added advantage of being tested :)

llvm-svn: 26243
```
  936cc9fe
- Dumb bug. Code sees a memcpy from X+c so it increments src offset. But it · c3dcf5a4
  Evan Cheng authored Feb 16, 2006
```
turns out not to point to a constant string but it forgot change the offset
back.

llvm-svn: 26242
```
  c3dcf5a4
Feb 16, 2006

1. Use pxor instead of xoraps / xorapd to clear FR32 / FR64 registers. This · 24c461b5

Evan Cheng authored Feb 16, 2006

   proves to be worth 20% on Ptrdist/ks. Might be related to dependency
   breaking support.
2. Added FsMOVAPSrr and FsMOVAPDrr as aliases to MOVAPSrr and MOVAPDrr. These
   are used for FR32 / FR64 reg-to-reg copies.
3. Tell reg-allocator to generate MOVSSrm / MOVSDrm and MOVSSmr / MOVSDmr to
   spill / restore FsMOVAPSrr and FsMOVAPDrr.

llvm-svn: 26241

24c461b5