search for: scalarevolutionaliasanalysis

...see several possible solutions: 1. Add yet-another hack to BasicAA to catch this case. 2. Let LICM use SE directly. Instead of bailing out on hoisting a loop-invariant value in an alias set with non-invariant values, use SE to check independence with the other values in the alias set. 3. Use ScalarEvolutionAliasAnalysis; scheduling SCEV AA prior to running LICM fixes this problem. Using #3 seems like the simplest, and most natural, way to fix this issue. I recall someone tell me (maybe at the developers' meeting) that SCEV AA was 'fundamentally flawed', and now I'm wondering what was meant by that...

...TypeBasedAliasAnalysis before BasicAliasAnalysis so that // BasicAliasAnalysis wins if they disagree. This is intended to help // support "obvious" type-punning idioms. PM.add(createTypeBasedAliasAnalysisPass()); PM.add(createBasicAliasAnalysisPass()); } ------ My goal is to use ScalarEvolutionAliasAnalysis in MemoryDependenceAnalysis. When I do, getPassName in MemoryDependenceAnalysis (by stepping into getAnalysisUsage), I get $39 = 0x399f778 "Basic Alias Analysis (stateless AA impl)" To switch to ScalarEvolutionAnalysis, I changed lib/Transforms/IPO/PassManagerBuilder.cpp: addInitialAli...

...mit-llvm -S ptr_test.c -o ptr_test.O1.ll` Crashed Command `opt -S -disable-output -basicaa -scev-aa -aa-eval -print-no-aliases ptr_test.O1.ll` The core dump message is also in the attachments. Investigations 1. SCEV-AA try to ‘minus' the SCEV expressions of the given two pointers(lib/Analysis/ScalarEvolutionAliasAnalysis.cpp:64) 2. ScalarEvolution::getMinusSCEV will boil down into ScalarEvolution::getAddExpr. On line 2383 of lib/Analysis/ScalarEvolution.cpp, GroupByComplexity is invoked. 3. CompareSCEVComplexity is eventually called to give a partial order between two SCEV expression. 4. If there are two SCEVAddExp...

...ions: > > 1. Add yet-another hack to BasicAA to catch this case. > > 2. Let LICM use SE directly. Instead of bailing out on hoisting a loop-invariant value in an alias set with non-invariant values, use SE to check independence with the other values in the alias set. > > 3. Use ScalarEvolutionAliasAnalysis; scheduling SCEV AA prior to running LICM fixes this problem. > > Using #3 seems like the simplest, and most natural, way to fix this issue. I recall someone tell me (maybe at the developers' meeting) that SCEV AA was 'fundamentally flawed', and now I'm wondering what was mea...

...1. Add yet-another hack to BasicAA to catch this case. >> >> 2. Let LICM use SE directly. Instead of bailing out on hoisting a loop-invariant value in an alias set with non-invariant values, use SE to check independence with the other values in the alias set. >> >> 3. Use ScalarEvolutionAliasAnalysis; scheduling SCEV AA prior to running LICM fixes this problem. >> >> Using #3 seems like the simplest, and most natural, way to fix this issue. I recall someone tell me (maybe at the developers' meeting) that SCEV AA was 'fundamentally flawed', and now I'm wondering what...

...do not change, but that "a and b" are pointer equal > always. > > > What I describe is how, AFAICT, it is implemented in all of the alias > passes. > For example, SCEVAA, which actually computes the recurrences for these > loop expressions, has this: > // This is ScalarEvolutionAliasAnalysis. Get the SCEVs! > const SCEV *AS = SE->getSCEV(const_cast<Value *>(LocA.Ptr)); > const SCEV *BS = SE->getSCEV(const_cast<Value *>(LocB.Ptr)); > > // If they evaluate to the same expression, it's a MustAlias. > if (AS == BS) return MustAlias; > &gt...

...> > `opt -S -disable-output -basicaa -scev-aa -aa-eval -print-no-aliases > ptr_test.O1.ll` > > The core dump message is also in the attachments. > > > > Investigations > > 1. SCEV-AA try to ‘minus' the SCEV expressions of the given two > pointers(lib/Analysis/ScalarEvolutionAliasAnalysis.cpp:64) > > 2. ScalarEvolution::getMinusSCEV will boil down into > ScalarEvolution::getAddExpr. On line 2383 of > lib/Analysis/ScalarEvolution.cpp, GroupByComplexity is invoked. > > 3. CompareSCEVComplexity is eventually called to give a partial order > between two SCEV express...

Hello all, I've read the documentation on alias analysis, and I think I understand it literally, but I just want to be sure, because it seems a bit strange. As it says on this web page <http://llvm.org/docs/AliasAnalysis.html>, The MayAlias response is used whenever the two pointers might refer to the > same object. > > The PartialAlias response is used when the two memory

...alysis so that > // BasicAliasAnalysis wins if they disagree. This is intended to help > // support "obvious" type-punning idioms. > PM.add(createTypeBasedAliasAnalysisPass()); > PM.add(createBasicAliasAnalysisPass()); > } > ------ > > My goal is to use ScalarEvolutionAliasAnalysis in MemoryDependenceAnalysis. > When I do, getPassName in MemoryDependenceAnalysis (by stepping into > getAnalysisUsage), I get > $39 = 0x399f778 "Basic Alias Analysis (stateless AA impl)" > > To switch to ScalarEvolutionAnalysis, I changed > lib/Transforms/IPO/PassManag...

...A to catch this case. > > > > 2. Let LICM use SE directly. Instead of bailing out on hoisting a > > loop-invariant value in an alias set with non-invariant values, > > use SE to check independence with the other values in the alias > > set. > > > > 3. Use ScalarEvolutionAliasAnalysis; scheduling SCEV AA prior to > > running LICM fixes this problem. > > > > Using #3 seems like the simplest, and most natural, way to fix this > > issue. I recall someone tell me (maybe at the developers' meeting) > > that SCEV AA was 'fundamentally flawed'...

...gt; always. > >> > >> > >> What I describe is how, AFAICT, it is implemented in all of the alias > >> passes. > >> For example, SCEVAA, which actually computes the recurrences for these > >> loop expressions, has this: > >> // This is ScalarEvolutionAliasAnalysis. Get the SCEVs! > >> const SCEV *AS = SE->getSCEV(const_cast<Value *>(LocA.Ptr)); > >> const SCEV *BS = SE->getSCEV(const_cast<Value *>(LocB.Ptr)); > >> > >> // If they evaluate to the same expression, it's a MustAlias. > >&gt...

...ease+Asserts build llvm[3]: Compiling X86ISelDAGToDAG.cpp for Release+Asserts build llvm[3]: Compiling DeadStoreElimination.cpp for Release+Asserts build llvm[3]: Compiling EarlyCSE.cpp for Release+Asserts build llvm[3]: Compiling SelectionDAGPrinter.cpp for Release+Asserts build llvm[2]: Compiling ScalarEvolutionAliasAnalysis.cpp for Release+Asserts build llvm[3]: Compiling FlattenCFGPass.cpp for Release+Asserts build llvm[3]: Compiling StatepointLowering.cpp for Release+Asserts build llvm[2]: Compiling ScalarEvolutionExpander.cpp for Release+Asserts build llvm[3]: Compiling X86ISelLowering.cpp for Release+Asserts build...