RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 01:03:52 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
5,367 Commits 4 Branches 5 Tags
Commit Graph

2990 Commits

Author SHA1 Message Date
Chris Lattner
c37308604f Change behavior of changeExitBlock function to replace all instances of exit block 
llvm-svn: 5661
 2003-02-27 22:37:44 +00:00
Chris Lattner
74b9aaa89c Add a new assertion to check that stuff is happening right 
Ironically the exit block modification code wasn't updating the exit block
information itself.  Fix this.

llvm-svn: 5659
 2003-02-27 22:31:07 +00:00
Chris Lattner
d40fa5ca2f Fix bug: 2003-02-27-StoreSinkPHIs.ll 
llvm-svn: 5658
 2003-02-27 21:59:36 +00:00
Chris Lattner
e1d4fb3c2c Fix bug: LICM/2003-02-27-PreheaderProblem.ll 
There may be a single outside predecessor and
still need a new loop-preheader if the predecessor has multiple
successors.

llvm-svn: 5656
 2003-02-27 21:51:38 +00:00
Chris Lattner
e8dda81b22 * Don't forget to update Loop information! 
* Remove bogus assertion: there may be a single outside predecessor and
    still need a new loop-preheader if the predecessor has multiple
    successors.  See bug: LICM/2003-02-27-PreheaderProblem.ll

llvm-svn: 5655
 2003-02-27 21:50:19 +00:00
Chris Lattner
3b3cce733d Replace assertion with a handler. 
llvm-svn: 5653
 2003-02-27 20:55:48 +00:00
Chris Lattner
b078b28fa5 * Significant changes to the preheader insertion pass: 
- Now we perform loop exit-block splitting to ensure exit blocks are
     always dominated by the loop header.
   - We now preserve dominance frontier information
   - This fixes bug: LICM/2003-02-26-LoopExitNotDominated.ll

llvm-svn: 5652
 2003-02-27 20:27:08 +00:00
Chris Lattner
ac72aab6e7 - LoopInfo now calculates and tracks loop exit blocks 
llvm-svn: 5650
 2003-02-27 00:38:34 +00:00
Chris Lattner
5baf8bb966 Make the aliassettracker much more precise by actually tracking size 
information for various accesses.  What a concept.

llvm-svn: 5647
 2003-02-26 22:11:00 +00:00
Chris Lattner
933b11a154 Fix bug: BasicAA/2003-02-26-AccessSizeTest.ll 
llvm-svn: 5645
 2003-02-26 21:57:23 +00:00
Chris Lattner
e923b037b7 Fix several bugs in basic-aa 
llvm-svn: 5643
 2003-02-26 21:28:49 +00:00
Chris Lattner
ea424f3aff Rename -no-* to -disable-* 
llvm-svn: 5642
 2003-02-26 20:00:41 +00:00
Chris Lattner
4e622afb23 Add new -no-aa implementation 
llvm-svn: 5641
 2003-02-26 19:57:10 +00:00
Chris Lattner
721f9fc770 Move BasicAA pass out to it's own header file 
llvm-svn: 5640
 2003-02-26 19:41:54 +00:00
Chris Lattner
4162adcfee Adjust to implement new AA interface 
llvm-svn: 5638
 2003-02-26 19:29:36 +00:00
Chris Lattner
4d57299049 Adjust to new AA interface 
llvm-svn: 5637
 2003-02-26 19:29:16 +00:00
Chris Lattner
92a8816086 Convert to work with new AliasAnalysis interface by conservatively assuming all pointers are arbitrarily large accesses 
llvm-svn: 5636
 2003-02-26 19:28:57 +00:00
Chris Lattner
5afbabc83b Calculate and pass load sizes to the alias analysis infrastructure 
llvm-svn: 5635
 2003-02-26 19:27:35 +00:00
Chris Lattner
d67ff79f57 Adjust to new AA interface 
Add tracking for Mod/Ref info

llvm-svn: 5634
 2003-02-26 19:27:05 +00:00
Chris Lattner
a84b1da40f - Checkin of the alias analysis work: 
* Takes into account the size of the memory reference to determine aliasing.
    * Expose mod/ref information in a more consistent way
    * BasicAA can now disambiguate A[i][1] and A[j][2] for conservative request
      sizes

llvm-svn: 5633
 2003-02-26 19:26:51 +00:00
Chris Lattner
b7d52c4fa8 Allow ImmutablePass's to require other immutable passes and to be initialized 
llvm-svn: 5630
 2003-02-26 19:10:57 +00:00
Chris Lattner
ce3e651af6 Fix typeo 
llvm-svn: 5624
 2003-02-25 21:14:59 +00:00
Chris Lattner
b1aeaf7efb Remove wierd case that can never happen 
llvm-svn: 5623
 2003-02-25 20:27:09 +00:00
Chris Lattner
497c0d443c Rename Instruction::hasSideEffects() -> mayWriteToMemory() 
llvm-svn: 5620
 2003-02-24 20:48:32 +00:00
Chris Lattner
f2009dbddb This is a substantial rewrite of the AliasSetTracker class which now uses 
a union-find based algorithm, is significantly faster, and is more general.
It will also scale to handle call instructions correctly, which is a nice
added bonus.

This includes a new pass -print-alias-sets which can be used to show how
alias sets are formed for a particular analysis.

llvm-svn: 5619
 2003-02-24 20:37:56 +00:00
Chris Lattner
1c1f11bd3f Initial implementation of Loop invariant memory->scalar promotion 
llvm-svn: 5613
 2003-02-24 03:52:32 +00:00
Chris Lattner
2e4c1c8640 Split mem2reg promotion into two parts: a function which does the work, and 
a pass which wraps the function.  This allows other passes to use the functionality

llvm-svn: 5610
 2003-02-22 23:57:48 +00:00
Chris Lattner
283d117dad Clean up std namespace references 
llvm-svn: 5608
 2003-02-22 22:25:17 +00:00
Chris Lattner
4c2012352d Dramatically simplify building of natural loops and fix a bug where the BBMap 
was not correctly computed.

llvm-svn: 5606
 2003-02-22 21:33:11 +00:00
Chris Lattner
1fc8307d2a Fix the requisite bug that I introduced 
llvm-svn: 5605
 2003-02-20 00:28:00 +00:00
Chris Lattner
92bf8c2754 Fix 80 character formatting 
llvm-svn: 5604
 2003-02-20 00:18:07 +00:00
Chris Lattner
97d323caea Fix bug: 2003-02-19-LoopInfoNestingBug.ll 
llvm-svn: 5603
 2003-02-20 00:17:17 +00:00
Chris Lattner
c24ffa9297 4 new transformations: 
* X*C + X --> X * (C+1)
  * X + X*C --> X * (C+1)
  * X - X*C --> X * (1-C)
  * X*C - X --> X * (C-1)

llvm-svn: 5592
 2003-02-18 19:57:07 +00:00
Chris Lattner
8306cc5c5c Add a variety of new transformations: 
* A & ~A == 0
  * A / (2^c) == A >> c  if unsigned
  * 0 / A == 0
  * 1.0 * A == A
  * A * (2^c) == A << c
  * A ^ ~A == -1
  * A | ~A == -1
  * 0 % X = 0
  * A % (2^c) == A & (c-1) if unsigned
  * A - (A & B) == A & ~B
  * -1 - A == ~A

llvm-svn: 5587
 2003-02-18 19:28:33 +00:00
Chris Lattner
148d67f088 Fix resolution of indirect function calls... whoops 
llvm-svn: 5576
 2003-02-14 23:27:18 +00:00
Anand Shukla
7aae29d426 Changes to runtime framework 
llvm-svn: 5572
 2003-02-14 20:41:53 +00:00
Chris Lattner
37735c4677 Mark function edges 
llvm-svn: 5571
 2003-02-14 20:25:47 +00:00
Chris Lattner
174a241cd5 Fix a bug that cause a crash resolving questionable function calls 
llvm-svn: 5567
 2003-02-14 19:12:29 +00:00
Chris Lattner
00fe161e49 - Eliminate provably non-pointer nodes from graphs. 
This helps a lot of testcases, for example:
            New Time  New #Nodes     Old Time    Old #Nodes
254.gap:    91.1024     21605        91.1397     22657
povray31:    2.7807      8613         3.0152     10338
255.vortex:  1.2034      8153         1.2172     8822
moria:        .6756      3150          .7054     3877
300.twolf:    .1652      2010          .1851     3270

Typically, testcases which use long and ulong integers a lot get better, f.e. povray above.

llvm-svn: 5566
 2003-02-14 06:28:00 +00:00
Chris Lattner
8de17c9baf Fix a misunderstanding of the standard associative containers 
llvm-svn: 5565
 2003-02-14 05:34:36 +00:00
Chris Lattner
e5dbced215 Don't put integer pointers (longs) into the scalar map. 
This speeds stuff up by 10% on some tests, woot!

llvm-svn: 5564
 2003-02-14 04:55:58 +00:00
Chris Lattner
e70bdb4355 Fix bug: LevelRaise/2003-02-13-CallRaise.ll 
llvm-svn: 5562
 2003-02-14 03:07:17 +00:00
Chris Lattner
b178c26c20 Hax0r around a deficiency in the Pass infrastructure 
llvm-svn: 5555
 2003-02-13 21:44:18 +00:00
Chris Lattner
32a8143e30 Use the new tailclip property new added to graphviz CVS to make call nodes 
actually intelligable

llvm-svn: 5554
 2003-02-13 20:14:40 +00:00
Chris Lattner
c4affd76ec This speeds up processing LLVM a _lot_, 17% in the case of loading and destroying "vortex" 
llvm-svn: 5553
 2003-02-13 19:46:22 +00:00
Chris Lattner
120d135f05 Move node forwarding code from being inlined to being out-of-line. 
This brings a 11.6% speedup to steens, and a 3.6 overall speedup to ds-aa

llvm-svn: 5552
 2003-02-13 19:09:00 +00:00
Chris Lattner
9efc110b12 Make more compatible with GCC 2.96 
llvm-svn: 5550
 2003-02-13 16:25:28 +00:00
Chris Lattner
34094ff7e3 Remove gunk that was supposed to make space evaluation more precise, but never worked. 
llvm-svn: 5549
 2003-02-13 05:07:53 +00:00
Joel Stanley
1b7b1786a5 Fixed alloca declaration problem on sun. 
llvm-svn: 5546
 2003-02-12 20:45:00 +00:00
Chris Lattner
bf2c5b238f Don't output times in "scientific" notation 
llvm-svn: 5544
 2003-02-12 19:02:13 +00:00