RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-25 04:02:41 +01:00
Code Issues Projects Releases Wiki Activity
7,207 Commits 4 Branches 5 Tags 704 MiB
25a9f3f3ba
Commit Graph

364 Commits

Author SHA1 Message Date
Misha Brukman
74cf3f5246 s/convertable/convertible/g 
llvm-svn: 6248
 2003-05-20 18:45:36 +00:00
Chris Lattner
d8234cfa3a Fix long standing bug 
llvm-svn: 6232
 2003-05-15 18:25:13 +00:00
Chris Lattner
6cb27840a7 Fix bug: LoopPreheaders/2003-05-12-PreheaderExitOfChild.ll 
llvm-svn: 6153
 2003-05-12 22:04:34 +00:00
Chris Lattner
a7ee27e0fd Handle va_arg instruction correctly 
llvm-svn: 6030
 2003-05-08 02:50:13 +00:00
Chris Lattner
e006ca4580 Fix spelling 
llvm-svn: 5983
 2003-05-02 19:26:34 +00:00
Chris Lattner
f13fe891d1 Fix Bug: LowerSwitch/2003-05-01-PHIProblem.ll 
llvm-svn: 5979
 2003-05-01 22:21:21 +00:00
Chris Lattner
5045ab0fac Fix another case where constexprs could cause a crash 
llvm-svn: 5972
 2003-04-30 22:34:06 +00:00
Chris Lattner
caa766b40f Fix constant folding of constexprs 
llvm-svn: 5971
 2003-04-30 22:19:10 +00:00
Chris Lattner
d2960ee802 Fix bug: ADCE/2003-04-25-PHIPostDominateProblem.ll 
llvm-svn: 5952
 2003-04-25 22:53:27 +00:00
Chris Lattner
c78cd5c758 Fix major problem with PHI node traversal: rechecking PHIs should go into 
the visit function.  The worklist is for when their lattice value changes.

llvm-svn: 5927
 2003-04-25 03:35:10 +00:00
Chris Lattner
7b67a6e77d Minor non-functional changes: 
* Spell propagate right
  * Improve performance of phi node handling
  * Delete using directive
  * Other minor changes

llvm-svn: 5920
 2003-04-25 02:50:03 +00:00
Chris Lattner
db6b7ba364 Remove unnecesary &*'s 
llvm-svn: 5872
 2003-04-23 16:37:45 +00:00
Chris Lattner
a9eaa10358 New pass to lower switch instructions to branch instructions 
llvm-svn: 5865
 2003-04-23 16:23:59 +00:00
Chris Lattner
d609c99c87 Minor cleanups 
llvm-svn: 5863
 2003-04-23 16:18:14 +00:00
Chris Lattner
bb00f755ed Change the interface to constant expressions to allow automatic folding 
llvm-svn: 5793
 2003-04-16 22:40:51 +00:00
Chris Lattner
731913bab3 Initial checkin of PRE on LLVM. This implementation is still lacking in 
several ways:

 * Load expressions are not PRE'd well.  Alias Analysis should be used to
   get accurate information when computing anticipatibility.
 * The expression collection implementation does not handle PHI nodes properly,
   thus the implementation misses many opportunities to PRE.
 * This code could be sped up quite a bit

Despite these flaws, the code seems to work well, and handles PR's as one
would expect.

llvm-svn: 5759
 2003-03-31 19:55:43 +00:00
Chris Lattner
b215db890f Move BreakCriticalEdges pass to lib/Transforms/Utils 
llvm-svn: 5754
 2003-03-21 21:43:19 +00:00
Chris Lattner
c51d7871ee Add the following instcombine xforms: 
- Implement simple reassociation: (A|c1)|(B|c2) == (A|B)|(c1|c2)
  - (A & C1)+(B & C2) -> (A & C1)|(B & C2) iff C1&C2 == 0
  - (A & C1)^(B & C2) -> (A & C1)|(B & C2) iff C1&C2 == 0

llvm-svn: 5743
 2003-03-11 00:12:48 +00:00
Chris Lattner
6eededc41f Implement: -A*-B == A*B 
llvm-svn: 5740
 2003-03-10 23:23:04 +00:00
Chris Lattner
715b592fc4 Add new transformation: // (~A | ~B) == (~(A & B)) 
llvm-svn: 5738
 2003-03-10 23:13:59 +00:00
Chris Lattner
0f5ab645eb Generalize not and neg comparison testers to allow constant to be considered not'able and neg'able. This 
allows optimization of this:
int %test4(int %A, int %B) {
        %a = xor int %A, -1
        %c = and int %a, 5    ; 5 = ~c2
        %d = xor int %c, -1
        ret int %d
}

into this:
int %test4(int %A, int %B) {            ; No predecessors!
        %c.demorgan = or int %A, -6             ; <int> [#uses=1]
        ret int %c.demorgan
}

llvm-svn: 5736
 2003-03-10 23:06:50 +00:00
Chris Lattner
3eca9aaa0a Generalize (A+c1)+c2 optimization to work with all associative operators 
llvm-svn: 5733
 2003-03-10 21:43:22 +00:00
Chris Lattner
9dec4368c0 Minor change, no functionality diff 
llvm-svn: 5731
 2003-03-10 19:20:30 +00:00
Chris Lattner
5fbfe9b9a8 Fix bug: (x << 100) wasn't folded to 0, but (x >> 100) was (when x is unsigned) 
Implement new shift optimizations for shifting the result of a shift.

llvm-svn: 5729
 2003-03-10 19:16:08 +00:00
Chris Lattner
5e6fb8b8d5 Implement: (A|B)^B == A & (~B) 
llvm-svn: 5728
 2003-03-10 18:24:17 +00:00
Chris Lattner
bd009a8938 Implement %test7 in InstCombine/getelementptr.ll 
llvm-svn: 5704
 2003-03-05 22:33:14 +00:00
Chris Lattner
4eaaa99b93 Convert LICM over to use AliasSetTracker. Besides being nicer, this automatically 
allows LICM to use access sizes to help alias analysis be more precise.

llvm-svn: 5693
 2003-03-03 23:32:45 +00:00
Chris Lattner
a49e96dda5 Change the mem2reg interface to accept a TargetData argument 
llvm-svn: 5685
 2003-03-03 17:25:18 +00:00
Chris Lattner
09c35df4e9 Fix bug: LICM/2003-02-28-PromoteDifferentType.ll 
llvm-svn: 5675
 2003-02-28 19:21:40 +00:00
Chris Lattner
dec7727a9f Fix bug I introduced yesterday :( 
llvm-svn: 5669
 2003-02-28 16:54:17 +00:00
Chris Lattner
e2aebf5906 Fix bug: LICM/2003-02-27-PreheaderExitNodeUpdate.ll 
llvm-svn: 5667
 2003-02-28 03:07:54 +00:00
Chris Lattner
d64f135965 Fix bug: 2003-02-27-PreheaderExitNodeUpdate.ll by updating exit node info 
llvm-svn: 5664
 2003-02-27 22:48:57 +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
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
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
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
497c0d443c Rename Instruction::hasSideEffects() -> mayWriteToMemory() 
llvm-svn: 5620
 2003-02-24 20:48:32 +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
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
aceb57a289 Fix typo 
llvm-svn: 5458
 2003-02-01 04:50:59 +00:00
Chris Lattner
7af7b1c82c Fix bug: ADCE/2003-01-22-PredecessorProblem.ll 
llvm-svn: 5405
 2003-01-23 02:12:18 +00:00
Chris Lattner
61d594690a Fix a huge performance problem in reassociate by introducing a 
rank map cache for instruction ranks

llvm-svn: 5030
 2002-12-15 03:56:00 +00:00
Chris Lattner
18fe81719d Minor changes: 
* Reword comment to make more clear
  * Don't print out BB's after modification made
  * Don't delete and new an instruction when we need to move something, just move it.

llvm-svn: 5029
 2002-12-15 03:49:50 +00:00
Chris Lattner
be53e88459 Remove dead code 
llvm-svn: 4952
 2002-12-07 18:42:13 +00:00
Chris Lattner
6c9d57c590 Fix bug: 2002-12-05-MissedConstProp.ll pointed out by Casey Carter 
llvm-svn: 4935
 2002-12-05 22:41:53 +00:00
Chris Lattner
b5027095e5 - Eliminated the deferred symbol table stuff in Module & Function, it really 
wasn't an optimization and it was causing lots of bugs.

llvm-svn: 4779
 2002-11-20 18:36:02 +00:00
Chris Lattner
0141841bc4 Fix warning 
llvm-svn: 4649
 2002-11-09 00:49:43 +00:00
Chris Lattner
f03f32d2e8 Fix spelling error 
llvm-svn: 4645
 2002-11-08 23:18:37 +00:00
Chris Lattner
aaf2fa34a5 Instcombine this away: 
%A = getelementptr [4 x int]* %M, long 0, long 0
        %B = getelementptr int* %A, long 2

llvm-svn: 4536
 2002-11-04 16:43:32 +00:00
Chris Lattner
ecddf3396f Add a transformation to turn: 
malloc Ty, C
int
  malloc [C x Ty], 1

llvm-svn: 4534
 2002-11-04 16:18:53 +00:00
Chris Lattner
b9fa84c1ed Make sure to _delete_ memory allocated by worklists 
llvm-svn: 4523
 2002-11-04 02:54:22 +00:00
Chris Lattner
b913f4e785 Fixes to the reassociate pass to make it respect dominance properties 
Huge thanks go to Casey Carter for writing this fix, reassociate is now
reoperational!

llvm-svn: 4471
 2002-10-31 17:12:59 +00:00
Chris Lattner
1ec9725ef8 BreakCriticalEdges should update dominance frontier information as well as 
other dominance stuff.  Patch contributed by Casey Carter

llvm-svn: 4457
 2002-10-31 02:44:36 +00:00
Misha Brukman
c9c721edf0 Fix spelling of `propagate'. 
llvm-svn: 4423
 2002-10-29 23:06:16 +00:00
Chris Lattner
b565d3ac5b - Rename AnalysisUsage::preservesAll to getPreservesAll & preservesCFG to 
setPreservesCFG to be less confusing.

llvm-svn: 4255
 2002-10-21 20:00:28 +00:00
Chris Lattner
d68e6d1718 - Fix bug: cee/2002-10-07-NoImmediateDominator.ll 
llvm-svn: 4081
 2002-10-08 21:53:51 +00:00
Chris Lattner
40df2d9d4c - Checkin LARGE number of Changes to CEE pass that will make it much more 
powerful, but that are largely disabled.  The basic idea here is that it
    is trying to forward branches across basic blocks that have PHI nodes in
    it, which are crucial to be able to handle cases like whet.ll.
    Unfortunately we are not updating SSA correctly, causing sim.c to die, and I
    don't have time to fix the regression now, so I must disable the
    functionality.

llvm-svn: 4077
 2002-10-08 21:34:15 +00:00
Chris Lattner
3cf20c3f14 Expose isCriticalEdge & SplitCriticalEdge methods from crit-edges pass 
llvm-svn: 4075
 2002-10-08 21:06:27 +00:00
Chris Lattner
b32252aa71 It is illegal for PHI nodes to have zero values, delete the code to handle them 
llvm-svn: 4071
 2002-10-08 17:07:39 +00:00
Chris Lattner
55c7ae3cdc Fold ashr -1, X into -1 
llvm-svn: 4070
 2002-10-08 16:16:40 +00:00
Chris Lattner
7cf9f6f4b1 Updates to work with recent Statistic's changes: 
* Renamed StatisticReporter.h/cpp to Statistic.h/cpp
    * Broke constructor to take two const char * arguments instead of one, so
      that indendation can be taken care of automatically.
    * Sort the list by pass name when printing
    * Make sure to print all statistics as a group, instead of randomly when
      the statistics dtors are called.
    * Updated ProgrammersManual with new semantics.

llvm-svn: 4002
 2002-10-01 22:38:41 +00:00
Chris Lattner
61e8de1c61 Fix bug in LICM that caused the previous big win. :( 
llvm-svn: 3980
 2002-09-29 22:26:07 +00:00
Chris Lattner
ef0e95bb42 Hoist the contents of Loops in depth first order in the dominator tree, 
rather than in random order.  This causes LICM to be DRAMATICALLY more
effective. For example, on bzip2.c, it is able to hoist 302 loads and
2380 total instructions, as opposed to 44/338 before.  This  obviously
makes other transformations much more powerful as well!

llvm-svn: 3978
 2002-09-29 21:46:09 +00:00
Chris Lattner
b731880895 Fix major bugs in dominator set & tree information updating 
llvm-svn: 3975
 2002-09-29 21:41:38 +00:00
Chris Lattner
0bbb8f6e06 - Further cleanups of LICM pass, remove extra work from previous implementation 
- Do not clone instructions then insert clone outside of loop.  Just move them.

llvm-svn: 3951
 2002-09-26 19:40:25 +00:00
Chris Lattner
0c66734ac5 Improve comments, doxygenize more 
llvm-svn: 3950
 2002-09-26 16:52:07 +00:00
Chris Lattner
21eb78d59d Clean up LICM significantly now that it is guaranteed to have loop preheaders 
llvm-svn: 3947
 2002-09-26 16:38:03 +00:00
Chris Lattner
9cdbd428b0 Change pass name to something sane 
llvm-svn: 3946
 2002-09-26 16:37:37 +00:00
Chris Lattner
fe9fba6f5b Loop invariant code motion now depends on the LoopPreheader pass. Dead code 
has not yet been removed.

llvm-svn: 3945
 2002-09-26 16:19:31 +00:00
Chris Lattner
f9676d651d - Cleanup break-crit-edges pass by making SplitCriticalEdge a member method. 
- break-crit-edges pass does not invalidate loop-preheader pass.

llvm-svn: 3944
 2002-09-26 16:18:51 +00:00
Chris Lattner
e5996b4c88 Checkin new loop-preheader insertion pass. 
llvm-svn: 3942
 2002-09-26 16:17:31 +00:00
Chris Lattner
7ab2bba3d3 Change LowerAllocations pass to 'require' TargetData instead of it being 
passed in.

llvm-svn: 3929
 2002-09-25 23:47:47 +00:00
Chris Lattner
7fafece8e0 Eliminate extraneous #include 
llvm-svn: 3916
 2002-09-24 21:18:40 +00:00
Chris Lattner
25ba5ac7ce - Do not expose Critical Edge breaking mechanics outside the BCE pass, thus 
removing it from Transforms/Local.h and from Transforms/Utils/*

llvm-svn: 3910
 2002-09-24 15:51:56 +00:00
Chris Lattner
1014aae962 Correlated Exprs pass now requires BCE pass instead of doing it manually 
llvm-svn: 3908
 2002-09-24 15:43:56 +00:00
Chris Lattner
49588769e0 - Expose passinfo from BreakCriticalEdges pass so that it may be "Required" 
by other passes.  Make BCE pass be in anonymous namespace now.

llvm-svn: 3907
 2002-09-24 15:43:12 +00:00
Chris Lattner
e62cdff96c Add new BreakCriticalEdges pass 
llvm-svn: 3903
 2002-09-24 00:08:39 +00:00
Chris Lattner
b6c408778a Optimize away cases like: 
%cast109 = cast uint %cast212 to short          ; <short> [#uses=1]
        %cast214 = cast short %cast109 to uint          ; <uint> [#uses=1]
        %cast215 = cast uint %cast214 to short          ; <short> [#uses=1]

llvm-svn: 3897
 2002-09-23 23:39:43 +00:00
Chris Lattner
2a53929c5d * Fix bug: CorrelatedExprs/2002-09-23-PHIUpdateBug.ll 
* Make sure "Changed" is updated correctly

llvm-svn: 3891
 2002-09-23 20:06:22 +00:00
Chris Lattner
5ccb21f41d Fix bug: test/Regression/Transforms/InstCombine/2002-09-17-GetElementPtrCrash.ll 
Thanks to Casey for finding it!

llvm-svn: 3783
 2002-09-17 21:05:42 +00:00
Vikram S. Adve
4bcec53ed7 Extract most of the transformation into an externally accessible 
function -- DecomposeArrayRef(GetElementPtrInst* GEP) -- that can
be invoked on a single instruction at a time.

llvm-svn: 3755
 2002-09-16 16:40:07 +00:00
Misha Brukman
2e598f0092 Function.h is unnecessary when Module.h is included. 
llvm-svn: 3716
 2002-09-14 03:04:02 +00:00
Chris Lattner
fb017cb38c - Change getelementptr instruction to use long indexes instead of uint 
indexes for sequential types.

llvm-svn: 3682
 2002-09-11 01:21:33 +00:00
Chris Lattner
5f4b79edaf Fix the last of the silly bugs I just introduced. :( 
llvm-svn: 3674
 2002-09-10 23:46:10 +00:00
Chris Lattner
5143d4a3b3 Add cannonicalization of shl X, 1 -> add X, X 
llvm-svn: 3671
 2002-09-10 23:04:09 +00:00
Chris Lattner
37c3a1d80c Clean up code due to auto-insert constructors 
llvm-svn: 3665
 2002-09-10 22:38:47 +00:00
Chris Lattner
da9702a8a1 Simplify code (somtimes dramatically), by using the new "auto-insert" feature 
of instruction constructors.

llvm-svn: 3656
 2002-09-10 17:04:02 +00:00
Chris Lattner
7787814c22 * Clean up code a little bit 
* Fix bug: test/Regression/Transforms/IndVarsSimplify/2002-09-09-PointerIndVar.ll

llvm-svn: 3644
 2002-09-10 05:24:05 +00:00
Chris Lattner
4f5abce7f6 Hack unneccesary now that shifts of pointers are no longer legal! 
llvm-svn: 3640
 2002-09-10 03:50:54 +00:00
Chris Lattner
c1de581f8e Fix Bug: test/Regression/Transforms/InstCombine/2002-09-08-PointerShiftBug.ll 
llvm-svn: 3626
 2002-09-08 21:39:07 +00:00
Chris Lattner
c30029172a Minor simplification 
llvm-svn: 3619
 2002-09-08 18:55:04 +00:00
Chris Lattner
bd03f18e04 Initial checkin of Correlated Expression Elimination Pass 
llvm-svn: 3599
 2002-09-06 18:41:55 +00:00
Chris Lattner
3ce5b343c5 - Renamed Type::isIntegral() to Type::isInteger() 
- Added new method Type::isIntegral() that is the same as isInteger, but
    also accepts bool.

llvm-svn: 3574
 2002-09-03 01:08:28 +00:00
Chris Lattner
cd49f0ec5d Add constant prop & DIE to InstCombine, so it cleans up after itself 
llvm-svn: 3568
 2002-09-02 04:59:56 +00:00
Chris Lattner
bb8f1a072b - Implement SCCP of getelementptr instructions 
- Implement SCCP of globals into ConstantExprs hopefully opening new
   opportunities for more SCCP.

llvm-svn: 3555
 2002-08-30 23:39:00 +00:00
Chris Lattner
a146c9e769 Really minor cleanups 
llvm-svn: 3549
 2002-08-30 22:53:30 +00:00
Chris Lattner
b0cbcfd5e4 - GCSE now no longer counts instructions not removed (due to no common 
dominator as being removed)
  - GCSE now uses new Value #'ing interface, instead of dealing with AA itself
  - GCSE worklist implementation much simpler, class cleaned up.

llvm-svn: 3533
 2002-08-30 20:22:29 +00:00