summaryrefslogtreecommitdiffstats
path: root/llvm/lib/Transforms
Commit message (Collapse)AuthorAgeFilesLines
* Generalize OptimizeLoopTermCond to optimize more loop terminating icmp to ↵Evan Cheng2009-11-171-87/+99
| | | | | | use postinc iv. llvm-svn: 89116
* Remove trailing whitespaceJim Grosbach2009-11-171-74/+74
| | | | llvm-svn: 89110
* Remove debug info attached with an instruction.Devang Patel2009-11-171-36/+18
| | | | llvm-svn: 89016
* Fix an expensive-checks error.David Greene2009-11-161-20/+24
| | | | | | | The Mask and LHSMask may not be of the same size, so don't do the transformation if they're different. llvm-svn: 88972
* CreateIntCast takes an "isSigned" parameter. Pass "true" for it, rather thanDuncan Sands2009-11-161-1/+3
| | | | | | a name. llvm-svn: 88908
* make PRE of loads preserve the alignment of the moved load instruction.Chris Lattner2009-11-151-5/+8
| | | | llvm-svn: 88865
* fix a bug handling 'not x' when x is undef.Chris Lattner2009-11-151-2/+3
| | | | llvm-svn: 88864
* Revert r88830 and r88831 which appear to have caused a selfhost buildbot someNick Lewycky2009-11-151-28/+0
| | | | | | grief. I suspect this patch merely exposed a bug else. llvm-svn: 88841
* Teach instcombine to look for booleans in wider integers when it encounters aNick Lewycky2009-11-151-0/+28
| | | | | | | zext(icmp). It may be able to optimize that away. This fixes one of the cases in PR5438. llvm-svn: 88830
* Remove LLVMContext from reassociate. It was threaded through every function butNick Lewycky2009-11-141-22/+14
| | | | | | ultimately never used. llvm-svn: 88763
* Add an option for running GVN with redundant load processing disabled.Dan Gohman2009-11-141-9/+23
| | | | llvm-svn: 88742
* Re-enable this code, since redundant PHIs are now being better nuked.Owen Anderson2009-11-121-4/+0
| | | | llvm-svn: 87042
* use isInstructionTriviallyDead, as pointed out by DuncanChris Lattner2009-11-121-3/+3
| | | | llvm-svn: 87035
* implement a nice little efficiency hack in the inliner. Since we're nowChris Lattner2009-11-121-14/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | running IPSCCP early, and we run functionattrs interlaced with the inliner, we often (particularly for small or noop functions) completely propagate all of the information about a call to its call site in IPSSCP (making a call dead) and functionattrs is smart enough to realize that the function is readonly (because it is interlaced with inliner). To improve compile time and make the inliner threshold more accurate, realize that we don't have to inline dead readonly function calls. Instead, just delete the call. This happens all the time for C++ codes, here are some counters from opt/llvm-ld counting the number of times calls were deleted vs inlined on various apps: Tramp3d opt: 5033 inline - Number of call sites deleted, not inlined 24596 inline - Number of functions inlined llvm-ld: 667 inline - Number of functions deleted because all callers found 699 inline - Number of functions inlined 483.xalancbmk opt: 8096 inline - Number of call sites deleted, not inlined 62528 inline - Number of functions inlined llvm-ld: 217 inline - Number of allocas merged together 2158 inline - Number of functions inlined 471.omnetpp: 331 inline - Number of call sites deleted, not inlined 8981 inline - Number of functions inlined llvm-ld: 171 inline - Number of functions deleted because all callers found 629 inline - Number of functions inlined Deleting a call is much faster than inlining it, and is insensitive to the size of the callee. :) llvm-svn: 86975
* - Teach LSR to avoid changing cmp iv stride if it will create an immediate thatEvan Cheng2009-11-121-209/+355
| | | | | | | | | | | | | | cannot be folded into target cmp instruction. - Avoid a phase ordering issue where early cmp optimization would prevent the later count-to-zero optimization. - Add missing checks which could cause LSR to reuse stride that does not have users. - Fix a bug in count-to-zero optimization code which failed to find the pre-inc iv's phi node. - Remove, tighten, loosen some incorrect checks disable valid transformations. - Quite a bit of code clean up. llvm-svn: 86969
* use getPredicateOnEdge to fold comparisons through PHI nodes,Chris Lattner2009-11-121-1/+11
| | | | | | | which implements GCC PR18046. This also gets us 360 more jump threads on 176.gcc. llvm-svn: 86953
* various fixes to the lattice transfer functions.Chris Lattner2009-11-121-0/+6
| | | | llvm-svn: 86952
* switch jump threading to use getPredicateOnEdge in one placeChris Lattner2009-11-121-10/+13
| | | | | | | | making the new LVI stuff smart enough to subsume some special cases in the old code. Disable them when LVI is around, the testcase still passes. llvm-svn: 86951
* Add the braces gcc suggested.Daniel Dunbar2009-11-121-5/+8
| | | | llvm-svn: 86933
* with the new code we can thread non-instruction values. ThisChris Lattner2009-11-121-1/+6
| | | | | | allows us to handle the test10 testcase. llvm-svn: 86924
* this argument can be an arbitrary value, it doesn't need to be an instruction.Chris Lattner2009-11-121-4/+3
| | | | llvm-svn: 86923
* expose edge information and switch j-t to use it.Chris Lattner2009-11-121-2/+2
| | | | llvm-svn: 86920
* pass TD into a SimplifyCmpInst call. Add another case thatChris Lattner2009-11-111-3/+24
| | | | | | uses LVI info when -enable-jump-threading-lvi is passed. llvm-svn: 86886
* Don't trivially delete unused calls to llvm.invariant.start. This allowsDuncan Sands2009-11-111-0/+3
| | | | | | | | | | | | | llvm.invariant.start to be used without necessarily being paired with a call to llvm.invariant.end. If you run the entire optimization pipeline then such calls are in fact deleted (adce does it), but that's actually a good thing since we probably do want them to be zapped late in the game. There should really be an integration test that checks that the llvm.invariant.start call lasts long enough that all passes that do interesting things with it get to do their stuff before it is deleted. But since no passes do anything interesting with it yet this will have to wait for later. llvm-svn: 86840
* remove the now dead condprop pass, PR3906.Chris Lattner2009-11-113-294/+0
| | | | llvm-svn: 86810
* stub out some LazyValueInfo interfaces, and have JumpThreadingChris Lattner2009-11-111-18/+45
| | | | | | | | start using them in a trivial way when -enable-jump-threading-lvi is passed. enable-jump-threading-lvi will be my playground for awhile. llvm-svn: 86789
* add a fixmeChris Lattner2009-11-111-0/+4
| | | | llvm-svn: 86766
* Block terminator may be a switch.Evan Cheng2009-11-111-1/+1
| | | | llvm-svn: 86761
* Implement support to debug inlined functions.Devang Patel2009-11-102-4/+57
| | | | llvm-svn: 86748
* implement a TODO by teaching jump threading about "xor x, 1".Chris Lattner2009-11-101-2/+14
| | | | llvm-svn: 86739
* move some generally useful functions out of jump threadingChris Lattner2009-11-102-85/+47
| | | | | | into libanalysis and transformutils. llvm-svn: 86735
* fix a crash in SCCP handling extractvalue of an array, pointed out andChris Lattner2009-11-101-3/+8
| | | | | | tracked down by Stephan Reiter! llvm-svn: 86726
* improve comment.Chris Lattner2009-11-101-2/+5
| | | | llvm-svn: 86723
* Make jump threading eliminate blocks that just contain phi nodes,Chris Lattner2009-11-101-0/+21
| | | | | | | | debug intrinsics, and an unconditional branch when possible. This reuses the TryToSimplifyUncondBranchFromEmptyBlock function split out of simplifycfg. llvm-svn: 86722
* Generalize lsr code that optimize loop to count down towards zero.Evan Cheng2009-11-101-80/+126
| | | | llvm-svn: 86715
* Add defensive break.Duncan Sands2009-11-101-0/+1
| | | | llvm-svn: 86705
* Fix obvious typo.Duncan Sands2009-11-101-1/+1
| | | | llvm-svn: 86694
* clarify logic.Chris Lattner2009-11-101-4/+4
| | | | llvm-svn: 86689
* Teach DSE to eliminate useless trampolines.Duncan Sands2009-11-101-6/+19
| | | | llvm-svn: 86683
* Add brackets to make gcc-4.4 happy.Duncan Sands2009-11-101-3/+3
| | | | llvm-svn: 86681
* Update computeArraySize() to use ComputeMultiple() to determine the array ↵Victor Hernandez2009-11-101-7/+7
| | | | | | size associated with a malloc; also extend PerformHeapAllocSRoA() to check if the optimized malloc's arg had its highest bit set, so that it is safe for ComputeMultiple() to look through sext instructions while determining the optimized malloc's array size llvm-svn: 86676
* unify the code that determines whether it is a good idea to change the typeChris Lattner2009-11-101-27/+36
| | | | | | | of a computation. This fixes some infinite loops when dealing with TD that has no native types. llvm-svn: 86670
* Simplify.Nick Lewycky2009-11-101-3/+1
| | | | llvm-svn: 86668
* Reapply r86359, "Teach dead store elimination that certain intrinsics write toNick Lewycky2009-11-101-63/+118
| | | | | | | memory just like a store" with bug fixed (partial-overwrite.ll is the regression test). llvm-svn: 86667
* refactor TryToSimplifyUncondBranchFromEmptyBlock out of SimplifyCFG.Chris Lattner2009-11-102-164/+173
| | | | llvm-svn: 86666
* CMake: Support for building llvm loadable modules.Oscar Fuentes2009-11-101-1/+1
| | | | llvm-svn: 86656
* make jump threading recursively simplify expressions instead of doing it Chris Lattner2009-11-101-20/+42
| | | | | | | | | | | | | | | | | | | just one level deep. On the testcase we go from getting this: F1: ; preds = %T2 %F = and i1 true, %cond ; <i1> [#uses=1] br i1 %F, label %X, label %Y to a fully threaded: F1: ; preds = %T2 br label %Y This changes gets us to the point where we're forming (too many) switch instructions on doug's strswitch testcase. llvm-svn: 86646
* don't invalidate PN, rewrite of this code is in progress anyway.Chris Lattner2009-11-101-4/+5
| | | | llvm-svn: 86639
* add a new SimplifyInstruction API, which is like ConstantFoldInstruction, Chris Lattner2009-11-101-6/+5
| | | | | | | | except that the result may not be a constant. Switch jump threading to use it so that it gets things like (X & 0) -> 0, which occur when phi preds are deleted and the remaining phi pred was a zero. llvm-svn: 86637
* Fix DenseMap iterator constness.Jeffrey Yasskin2009-11-103-6/+6
| | | | | | | | | | | | | | | | | | | This patch forbids implicit conversion of DenseMap::const_iterator to DenseMap::iterator which was possible because DenseMapIterator inherited (publicly) from DenseMapConstIterator. Conversion the other way around is now allowed as one may expect. The template DenseMapConstIterator is removed and the template parameter IsConst which specifies whether the iterator is constant is added to DenseMapIterator. Actually IsConst parameter is not necessary since the constness can be determined from KeyT but this is not relevant to the fix and can be addressed later. Patch by Victor Zverovich! llvm-svn: 86636
OpenPOWER on IntegriCloud