summaryrefslogtreecommitdiffstats
path: root/llvm/test/Analysis/ScalarEvolution
Commit message (Collapse)AuthorAgeFilesLines
...
* Test case and comment for PR9633.Andrew Trick2011-04-271-0/+33
| | | | llvm-svn: 130294
* Propagate SCEV no-wrap flags whenever possible.Andrew Trick2011-03-153-14/+21
| | | | | | This needs review. llvm-svn: 127638
* When SCEV can determine the loop test is X < X, set ExactBECount=0.Andrew Trick2011-03-091-0/+34
| | | | | | | | When ExactBECount is a constant, use it for MaxBECount. When MaxBECount cannot be computed, replace it with ExactBECount. Fixes PR9424. llvm-svn: 127342
* teach SCEV that the scale and addition of an inbounds gep don't NSW.Chris Lattner2011-02-132-4/+31
| | | | | | | This fixes a FIXME in scev-aa.ll (allowing a new no-alias result) and generally makes things more precise. llvm-svn: 125449
* Per discussion with Dan G, inbounds geps *certainly* can haveChris Lattner2011-02-111-2/+2
| | | | | | | | | | | unsigned overflow (e.g. "gep P, -1"), and while they can have signed wrap in theoretical situations, modelling an AddRec as not having signed wrap is going enough for any case we can think of today. In the future if this isn't enough, we can revisit this. Modeling them as having NUW isn't causing any known problems either FWIW. llvm-svn: 125410
* Simplify some code with no functionality change. Make the test a lot moreNick Lewycky2011-01-231-15/+19
| | | | | | robust against smarter optimizations, using the power of FileCheck. llvm-svn: 124081
* Use value ranges to fold ext(trunc) in SCEV when possible.Nick Lewycky2011-01-231-0/+34
| | | | llvm-svn: 124062
* Implement requiredTransitiveTobias Grosser2011-01-201-0/+24
| | | | | | | The PassManager did not implement the transitivity of requiredTransitive. This was unnoticed since 2006. llvm-svn: 123942
* Similarly, analyze truncate through multiply.Nick Lewycky2011-01-191-0/+8
| | | | llvm-svn: 123842
* Add a missed SCEV fold that is required to continue analyzing the IR producedNick Lewycky2011-01-191-1/+9
| | | | | | | | | | | by indvars through the scev expander. trunc(add x, y) --> add(trunc x, y). Currently SCEV largely folds the other way which is probably wrong, but preserved to minimize churn. Instcombine doesn't do this fold either, demonstrating a missed optz'n opportunity on code doing add+trunc+add. llvm-svn: 123838
* Add a missing SCEV simplification sext(zext x) --> zext x.Nick Lewycky2011-01-191-0/+8
| | | | llvm-svn: 123832
* Revert the testcase from the previous reverted commit.Eric Christopher2011-01-111-2/+0
| | | | llvm-svn: 123227
* add a testcase I missed in previous commit.Chris Lattner2011-01-091-0/+2
| | | | llvm-svn: 123143
* teach SCEV analysis of PHI nodes that PHI recurences formedChris Lattner2011-01-091-2/+40
| | | | | | | with GEP instructions are always NUW, because PHIs cannot wrap the end of the address space. llvm-svn: 123105
* reduce indentation. Print <nuw> and <nsw> when dumping SCEV AddRec'sChris Lattner2011-01-091-1/+1
| | | | | | that have the bit set. llvm-svn: 123104
* remove some noise from tests.Chris Lattner2010-09-021-1/+1
| | | | llvm-svn: 112889
* When analyzing loop exit conditions combined with and and or, don'tDan Gohman2010-08-111-2/+40
| | | | | | | make any assumptions about when the two conditions will agree on when to permit the loop to exit. This fixes PR7845. llvm-svn: 110758
* Revert the part of r107257 which introduced new logic for usingDan Gohman2010-06-301-3/+4
| | | | | | | nsw and nuw flags from IR Instructions. On further consideration, this isn't valid. llvm-svn: 107298
* Add a testcase for scev-aa's new capability.Dan Gohman2010-06-301-3/+24
| | | | llvm-svn: 107258
* Allow "exhaustive" trip count evaluation on phi nodes with allDan Gohman2010-06-221-0/+19
| | | | | | constant operands. llvm-svn: 106537
* Fix ScalarEvolution's "exhaustive" trip count evaluation code to avoidDan Gohman2010-06-191-0/+31
| | | | | | | | assuming that loops are in canonical form, as ScalarEvolution doesn't depend on LoopSimplify itself. Also, with indirectbr not all loops can be simplified. This fixes PR7416. llvm-svn: 106389
* Revert r106304 (105548 and friends), which are the SCEVComplexityCompareDan Gohman2010-06-181-1/+1
| | | | | | optimizations. There is still some nondeterminism remaining. llvm-svn: 106306
* Reapply 105540, 105542, and 105548, and revert r105732.Dan Gohman2010-06-181-1/+1
| | | | llvm-svn: 106304
* Workaround SCEV non-determinism on this test, for now, to get buildbots back toDaniel Dunbar2010-06-091-1/+1
| | | | | | green. Dan, please revert this once the real problem is fixed. llvm-svn: 105732
* Optimize ScalarEvolution's SCEVComplexityCompare predicate: don't goDan Gohman2010-06-071-1/+1
| | | | | | | | | | | scrounging through SCEVUnknown contents and SCEVNAryExpr operands; instead just do a simple deterministic comparison of the precomputed hash data. Also, since this is more precise, it eliminates the need for the slow N^2 duplicate detection code. llvm-svn: 105540
* Add a comment to this test.Dan Gohman2010-04-261-0/+3
| | | | llvm-svn: 102387
* ScalarEvolution support for <= and >= loops.Dan Gohman2010-04-244-4/+31
| | | | | | | Also, generalize ScalarEvolutions's min and max recognition to handle some new forms of min and max that this change makes more common. llvm-svn: 102234
* Don't attempt to analyze values which are obviously undef. This fixes someDan Gohman2010-04-221-0/+39
| | | | | | assertion failures in extreme cases. llvm-svn: 102042
* Generalize ScalarEvolution's PHI analysis to handle loops that don'tDan Gohman2010-04-121-0/+29
| | | | | | | have preheaders or dedicated exit blocks, as clients may not otherwise need to run LoopSimplify. llvm-svn: 101030
* Avoid analyzing instructions in blocks not reachable from the entry block.Dan Gohman2010-03-091-0/+13
| | | | | | They are lots of trouble, and they don't matter. This fixes PR6559. llvm-svn: 98103
* Teach ScalarEvolution how to compute a tripcount for a loop withDan Gohman2010-02-191-0/+76
| | | | | | | | true or false as its exit condition. These are usually eliminated by SimplifyCFG, but the may be left around during a pass which wishes to preserve the CFG. llvm-svn: 96683
* -disable-output is no longer needed with -analyze.Dan Gohman2010-01-2653-54/+54
| | | | llvm-svn: 94574
* Fix the the ceiling-division used in computing the MaxBECount so that it doesn'tDan Gohman2010-01-262-2/+411
| | | | | | | | | | | | | have trouble with an intermediate add overflowing. Also, be more conservative about the case where the induction variable in an SLT loop exit can step past the RHS of the SLT and overflow in a single step. Make getSignedRange more aggressive, to recover for some common cases which the above fixes pessimized. This addresses rdar://7561161. llvm-svn: 94512
* Use WriteAsOperand instead of getName() to print loop header names,Dan Gohman2010-01-0918-27/+27
| | | | | | so that unnamed blocks are handled. llvm-svn: 93059
* Add a comment about A[i+(j+1)].Dan Gohman2009-12-011-1/+1
| | | | llvm-svn: 90185
* update status of this. basicaa is much improved now,Chris Lattner2009-11-261-2/+2
| | | | | | | only missing the one form (in this testcase). Dan, do you consider this example to be important? llvm-svn: 89953
* Make opt default to not adding a target data string and update tests that ↵Kenneth Uildriks2009-11-031-0/+1
| | | | | | depend on target data to supply it within the test llvm-svn: 85900
* Convert Analysis tests to FileCheck in regards to PR5307.Edward O'Callaghan2009-10-275-7/+17
| | | | llvm-svn: 85241
* Teach ScalarEvolution how to reason about no-wrap flags on loopsDan Gohman2009-09-171-0/+76
| | | | | | | | where the induction variable has a non-unit stride, such as {0,+,2}, and there are expressions such as {1,+,2} inside the loop formed with or or add nsw operators. llvm-svn: 82151
* Change tests from "opt %s" to "opt < %s" so that opt doesn't see theDan Gohman2009-09-1160-61/+61
| | | | | | | | input filename so that opt doesn't print the input filename in the output so that grep lines in the tests don't unintentionally match strings in the input filename. llvm-svn: 81537
* Use opt -S instead of piping bitcode output through llvm-dis.Dan Gohman2009-09-083-3/+3
| | | | llvm-svn: 81257
* Change these tests to feed the assembly files to opt directly, insteadDan Gohman2009-09-0860-61/+61
| | | | | | of using llvm-as, now that opt supports this. llvm-svn: 81226
* Create a ScalarEvolution-based AliasAnalysis implementation.Dan Gohman2009-08-261-0/+194
| | | | | | | | | | | | | | | | This is a simple AliasAnalysis implementation which works by making ScalarEvolution queries. ScalarEvolution has a more complete understanding of arithmetic than BasicAA's collection of ad-hoc checks, so it handles some cases that BasicAA misses, for example p[i] and p[i+1] within the same iteration of a loop. This is currently experimental. It may be that the main use for this pass will be to help find cases where BasicAA can be profitably extended, or to help in the development of the overall AliasAnalysis infrastructure, however it's also possible that it could grow up to become a directly useful pass. llvm-svn: 80098
* Loosen up the regex for this test so that it doesn't implicitlyDan Gohman2009-08-191-1/+3
| | | | | | depend on TargetData information. llvm-svn: 79491
* Make LLVM Assembly dramatically easier to read by aligning the comments,Dan Gohman2009-08-121-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | using formatted_raw_ostream's PadToColumn. Before: bb1: ; preds = %bb %2 = sext i32 %i.01 to i64 ; <i64> [#uses=1] %3 = getelementptr double* %p, i64 %2 ; <double*> [#uses=1] %4 = load double* %3, align 8 ; <double> [#uses=1] %5 = fmul double %4, 1.100000e+00 ; <double> [#uses=1] %6 = sext i32 %i.01 to i64 ; <i64> [#uses=1] %7 = getelementptr double* %p, i64 %6 ; <double*> [#uses=1] After: bb1: ; preds = %bb %2 = sext i32 %i.01 to i64 ; <i64> [#uses=1] %3 = getelementptr double* %p, i64 %2 ; <double*> [#uses=1] %4 = load double* %3, align 8 ; <double> [#uses=1] %5 = fmul double %4, 1.100000e+00 ; <double> [#uses=1] %6 = sext i32 %i.01 to i64 ; <i64> [#uses=1] %7 = getelementptr double* %p, i64 %6 ; <double*> [#uses=1] Several tests required whitespace adjustments. llvm-svn: 78816
* Change the assembly syntax for nsw, nuw, and exact, putting themDan Gohman2009-07-271-1/+1
| | | | | | | after their associated opcodes rather than before. This makes them a little easier to read. llvm-svn: 77194
* When attempting to sign-extend an addrec by interpretingDan Gohman2009-07-251-0/+74
| | | | | | | the step value as unsigned, the start value and the addrec itself still need to be treated as signed. llvm-svn: 77078
* Teach ScalarEvolution to make use of no-overflow flags whenDan Gohman2009-07-251-0/+40
| | | | | | analyzing add recurrences. llvm-svn: 77034
* Replace the original ad-hoc code for determining whether (v pred w) impliesDan Gohman2009-07-211-0/+37
| | | | | | | | (x pred y) with more thorough code that does more complete canonicalization before resorting to range checks. This helps it find more cases where the canonicalized expressions match. llvm-svn: 76671
* Add a testcase for PR4569, which is now fixed.Dan Gohman2009-07-211-0/+24
| | | | llvm-svn: 76526
OpenPOWER on IntegriCloud