Commits · 1ec5e73b328afbd57b5c954159251e876cd9ab71 · Roger Ferrer / llvm-epi-0.8

Jul 07, 2006
- Fix a typo that causes 2006-07-07-ComputeMaskedBits.ll to fail. · af5ae573
  Evan Cheng authored Jul 07, 2006
```
llvm-svn: 29072
```
  af5ae573
Jun 13, 2006
- Port some bugfixes in shift handling from SimplifyDemandedBits over to · c5bb8ab1
  Chris Lattner authored Jun 13, 2006
```
ComputeMaskedBits.  DemandedMasks and KnownZero/One masks should never have
bits set out of the range of the base datatype.

llvm-svn: 28768
```
  c5bb8ab1
- TargetLowering::ComputeMaskedBits was not clearing reciprocal bits on shifts. · 8cac9cd5
  Jim Laskey authored Jun 13, 2006
```
llvm-svn: 28765
```
  8cac9cd5
Jun 12, 2006
- Start on my todo list · 0e57b2cb
  Andrew Lenharth authored Jun 12, 2006
```
llvm-svn: 28752
```
  0e57b2cb
May 17, 2006
- Another typo. Pointed out by Nate Begeman. · 19aaaca2
  Evan Cheng authored May 17, 2006
```
llvm-svn: 28353
```
  19aaaca2
- Fix an obvious bug in getPackedTypeBreakdown. Return 1 if type is legal. · 6dcec44f
  Evan Cheng authored May 17, 2006
```
llvm-svn: 28351
```
  6dcec44f
May 16, 2006
- this should be 128 I think · 20eb2ce8
  Andrew Lenharth authored May 16, 2006
```
llvm-svn: 28330
```
  20eb2ce8
- Move this code to a common place · 1dc9ec58
  Andrew Lenharth authored May 16, 2006
```
llvm-svn: 28329
```
  1dc9ec58
May 12, 2006
- Refactor a bunch of includes so that TargetMachine.h doesn't have to include · 8c2c1e90
  Owen Anderson authored May 12, 2006
```
TargetData.h.  This should make recompiles a bit faster with my current
TargetData tinkering.

llvm-svn: 28238
```
  8c2c1e90
May 08, 2006
- When tracking demanded bits, if any bits from the sext of an SRA are demanded, · 10c65374
  Chris Lattner authored May 08, 2006
```
then so is the input sign bit.  This fixes mediabench/g721 on X86.

llvm-svn: 28166
```
  10c65374
May 07, 2006
- Use ComputeMaskedBits to determine # sign bits as a fallback. This allows us · cd4a6437
  Chris Lattner authored May 06, 2006
```
to handle all kinds of stuff, including silly things like:
sextinreg(setcc,i16) -> setcc.

llvm-svn: 28155
```
  cd4a6437
- Add some more sign propagation cases · 4f3de3e3
  Chris Lattner authored May 06, 2006
```
llvm-svn: 28154
```
  4f3de3e3
- Add some more simple sign bit propagation cases. · f8607573
  Chris Lattner authored May 06, 2006
```
llvm-svn: 28149
```
  f8607573
May 06, 2006
- Add some really really simple code for computing sign-bit propagation. · 7206d74f
  Chris Lattner authored May 06, 2006
```
This will certainly be enhanced in the future.

llvm-svn: 28145
```
  7206d74f
- Fold (trunc (srl x, c)) -> (srl (trunc x), c) · 86a1467f
  Chris Lattner authored May 06, 2006
```
llvm-svn: 28138
```
  86a1467f
- Implement ComputeMaskedBits/SimplifyDemandedBits for ISD::TRUNCATE · 0f64932a
  Chris Lattner authored May 05, 2006
```
llvm-svn: 28135
```
  0f64932a
May 03, 2006

Refactor TargetMachine, pushing handling of TargetData into the... · 20a631fd

Owen Anderson authored May 03, 2006

Refactor TargetMachine, pushing handling of TargetData into the target-specific subclasses. This has one caller-visible change: getTargetData() now returns a pointer instead of a reference.

This fixes PR 759.

llvm-svn: 28074

20a631fd

Apr 02, 2006
- relax assertion · 6c1321ca
  Chris Lattner authored Apr 02, 2006
```
llvm-svn: 27358
```
  6c1321ca
- Allow targets to compute masked bits for intrinsics. · e6025525
  Chris Lattner authored Apr 02, 2006
```
llvm-svn: 27357
```
  e6025525
Mar 31, 2006
- Was returning the wrong type. · 051f7861
  Chris Lattner authored Mar 31, 2006
```
llvm-svn: 27277
```
  051f7861
- Modify the TargetLowering::getPackedTypeBreakdown method to also return the · f144dac7
  Chris Lattner authored Mar 31, 2006
```
unpromoted element type.

llvm-svn: 27273
```
  f144dac7
- Implement TargetLowering::getPackedTypeBreakdown · 549fb167
  Chris Lattner authored Mar 31, 2006
```
llvm-svn: 27270
```
  549fb167
Mar 24, 2006
- Typo · 4b5b4e37
  Evan Cheng authored Mar 23, 2006
```
llvm-svn: 27008
```
  4b5b4e37
Mar 16, 2006
- set TransformToType correctly for vector types. · ad74844b
  Chris Lattner authored Mar 16, 2006
```
llvm-svn: 26797
```
  ad74844b
Mar 14, 2006
- Add LSR hooks. · af598d24
  Evan Cheng authored Mar 13, 2006
```
llvm-svn: 26740
```
  af598d24
Mar 13, 2006
- I can't convince myself that this is safe, remove the recursive call. · 3d761b62
  Chris Lattner authored Mar 13, 2006
```
llvm-svn: 26725
```
  3d761b62
Mar 05, 2006
- Do not fold (add (shl x, c1), (shl c2, c1)) -> (shl (add x, c2), c1), · c726a5c3
  Chris Lattner authored Mar 05, 2006
```
we want to canonicalize the other way.

llvm-svn: 26547
```
  c726a5c3
Mar 03, 2006
- Number of NodeTypes now exceeds 128. · a7fb285c
  Evan Cheng authored Mar 03, 2006
```
llvm-svn: 26503
```
  a7fb285c
Mar 01, 2006
- Add interfaces for targets to provide target-specific dag combiner optimizations. · 4a2eeea6
  Chris Lattner authored Mar 01, 2006
```
llvm-svn: 26442
```
  4a2eeea6
Feb 27, 2006

Implement bit propagation through sub nodes, this (re)implements · ab816404
Chris Lattner authored Feb 27, 2006
```
PowerPC/div-2.ll

llvm-svn: 26392
```
ab816404
Check RHS simplification before LHS simplification to avoid infinitely looping · a60751dd
Chris Lattner authored Feb 27, 2006
```
on PowerPC/small-arguments.ll

llvm-svn: 26389
```
a60751dd

Just like we use the RHS of an AND to simplify the LHS, use the LHS to · 27220f89

Chris Lattner authored Feb 27, 2006

simplify the RHS.  This allows for the elimination of many thousands of
ands from multisource, and compiles CodeGen/PowerPC/and-elim.ll:test2
into this:

_test2:
        srwi r2, r3, 1
        xori r3, r2, 40961
        blr

instead of this:

_test2:
        rlwinm r2, r3, 31, 17, 31
        xori r2, r2, 40961
        rlwinm r3, r2, 0, 16, 31
        blr

llvm-svn: 26388

27220f89

Add a bunch of missed cases. Perhaps the most significant of which is that · 118ddba9
Chris Lattner authored Feb 26, 2006
```
assertzext produces zero bits.

llvm-svn: 26386
```
118ddba9

Feb 24, 2006
- Recognize memory operand codes · 2a9e1e3e
  Chris Lattner authored Feb 24, 2006
```
llvm-svn: 26345
```
  2a9e1e3e
Feb 23, 2006
- Don't return registers from register classes that aren't legal. · 2e124af4
  Chris Lattner authored Feb 22, 2006
```
llvm-svn: 26317
```
  2e124af4
Feb 22, 2006
- split register class handling from explicit physreg handling. · 7ad77dfc
  Chris Lattner authored Feb 22, 2006
```
llvm-svn: 26308
```
  7ad77dfc
- Updates to match change of getRegForInlineAsmConstraint prototype · 7bb4696d
  Chris Lattner authored Feb 21, 2006
```
llvm-svn: 26305
```
  7bb4696d
Feb 18, 2006
- 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
Feb 17, 2006
- Fix bug noticed by VC++. · 0d62ebd1
  Jeff Cohen authored Feb 17, 2006
```
llvm-svn: 26252
```
  0d62ebd1
Feb 16, 2006

Rework the SelectionDAG-based implementations of SimplifyDemandedBits · 8a77efe4

Nate Begeman authored Feb 16, 2006

and ComputeMaskedBits to match the new improved versions in instcombine.
Tested against all of multisource/benchmarks on ppc.

llvm-svn: 26238

8a77efe4