search for: runonloop

Are LoopPasses allowed to maintain state between runOnLoop invocations? The documentation is not clear on this point. -Dave

On Aug 19, 2008, at 12:32 PM, David Greene wrote: > Are LoopPasses allowed to maintain state between runOnLoop > invocations? What kind of information are you looking to maintain between runOnLoop invocations ? > > The documentation is not clear on this point. doInitialization() and doFinalization() will be run for each pass before and after runOnLoop(). For each pass doInitialization(...

On Tuesday 19 August 2008 15:15, Devang Patel wrote: > On Aug 19, 2008, at 12:32 PM, David Greene wrote: > > Are LoopPasses allowed to maintain state between runOnLoop > > invocations? > > What kind of information are you looking to maintain between runOnLoop > invocations ? Cached analysis answers. > > The documentation is not clear on this point. > > doInitialization() and doFinalization() will be run for each pass > before and a...

...e output not reflect my changes? Index: lib/Transforms/Scalar/IndVarSimplify.cpp =================================================================== RCS file: /var/cvs/llvm/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp,v retrieving revision 1.78 diff -r1.78 IndVarSimplify.cpp 310c310 < runOnLoop(*I); --- > runOnLoop(LI, *I); 322,323c322,323 < void runOnLoop(Loop *L); < void EliminatePointerRecurrence(PHINode *PN, BasicBlock *Preheader, --- > void runOnLoop(LoopInfo *LI, Loop *L); > void EliminatePointerRecurrence(LoopInfo *LI, Loop *L, PHINode *PN, B...

On 12/12/11 9:59 AM, neda 8664 wrote: > hi all, > > I want access to all basic blocks of function in a loop, so I used the > following code: > > /bool parallel::runOnLoop(Loop *L, LPPassManager &LPM) > { > for (Function::iterator bi= func->begin(); bi != func->end(); bi++){ > // > } > }/ Are you modifying anything within this code (especially changes that add/remove basic blocks or change their Terminator instructions)? D...

hi all, I want access to all basic blocks of function in a loop, so I used the following code: *bool parallel::runOnLoop(Loop *L, LPPassManager &LPM) { for (Function::iterator bi= func->begin(); bi != func->end(); bi++){ // } }* First loop run without problem, but for second loop I get the following error: *0 libLLVM-2.9.so 0x0137d530 1 libLLVM-2.9.so 0x0137fa6c 2 0x002a7...

John Criswell <criswell <at> illinois.edu> writes: > > > On 12/12/11 9:59 AM, neda 8664 wrote: > > > hi all, > I want access to all basic blocks of function in a loop, so I used > the following code:bool parallel::runOnLoop(Loop *L, LPPassManager > &LPM) > { >     for (Function::iterator bi= func->begin(); bi != > func->end(); bi++){ >        // >      } > } > > > Are you modifying anything within th...

...2, 2011 at 7:42 PM, John Criswell <criswell at illinois.edu>wrote: > >> On 12/12/11 9:59 AM, neda 8664 wrote: >> >> hi all, >> >> I want access to all basic blocks of function in a loop, so I used the >> following code: >> >> *bool parallel::runOnLoop(Loop *L, LPPassManager &LPM) >> { >> for (Function::iterator bi= func->begin(); bi != func->end(); bi++){ >> // >> } >> }* >> >> >> Are you modifying anything within this code (especially changes that >> add/remove basi...

...mailto:criswell at illinois.edu>> wrote: > > On 12/12/11 9:59 AM, neda 8664 wrote: >> hi all, >> >> I want access to all basic blocks of function in a loop, so I >> used the following code: >> >> /bool parallel::runOnLoop(Loop *L, LPPassManager &LPM) >> { >> for (Function::iterator bi= func->begin(); bi != >> func->end(); bi++){ >> // >> } >> }/ > > Are you modifying anything within this code (...

...t;criswell at illinois.edu>wrote: >> >>> On 12/12/11 9:59 AM, neda 8664 wrote: >>> >>> hi all, >>> >>> I want access to all basic blocks of function in a loop, so I used the >>> following code: >>> >>> *bool parallel::runOnLoop(Loop *L, LPPassManager &LPM) >>> { >>> for (Function::iterator bi= func->begin(); bi != func->end(); bi++){ >>> // >>> } >>> }* >>> >>> >>> Are you modifying anything within this code (especially chan...

...> Index: lib/Transforms/Scalar/IndVarSimplify.cpp > =================================================================== > RCS file: /var/cvs/llvm/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp,v > retrieving revision 1.78 > diff -r1.78 IndVarSimplify.cpp > 310c310 > < runOnLoop(*I); > --- >> runOnLoop(LI, *I); > 322,323c322,323 > < void runOnLoop(Loop *L); > < void EliminatePointerRecurrence(PHINode *PN, BasicBlock *Preheader, > --- >> void runOnLoop(LoopInfo *LI, Loop *L); >> void EliminatePointerRecurrence(L...

I'm wondering if the documentation for LoopPass ( http://llvm.org/docs/WritingAnLLVMPass.html#LoopPass) is misleading or incorrect (or if I'm just missing something.) The documentation states: "The doFinalization method ... is called when the pass framework has finished calling runOnLoop<http://llvm.org/docs/WritingAnLLVMPass.html#runOnLoop> for every loop in the program being compiled." >From this, I understood that for a single instance of a LoopPass sub-class, the doFinalization method would be called exactly once, after all loops had been seen. However in practic...

...<http://llvm.org/docs/WritingAnLLVMPass.html#LoopPass>) >> is misleading or >> incorrect (or if I'm just missing something.) The documentation states: >> >> "The doFinalization method ... is called when the pass framework has >> finished >> calling runOnLoop <http://llvm.org/docs/**WritingAnLLVMPass.html#** >> runOnLoop <http://llvm.org/docs/WritingAnLLVMPass.html#runOnLoop>> for >> >> every loop in the program being compiled." >> >> From this, I understood that for a single instance of a LoopPass >&gt...

...g if the documentation for LoopPass > (http://llvm.org/docs/WritingAnLLVMPass.html#LoopPass) is misleading or > incorrect (or if I'm just missing something.) The documentation states: > > "The doFinalization method ... is called when the pass framework has finished > calling runOnLoop <http://llvm.org/docs/WritingAnLLVMPass.html#runOnLoop> for > every loop in the program being compiled." > > From this, I understood that for a single instance of a LoopPass sub-class, the > doFinalization method would be called exactly once, after all loops had been > se...

...;http://llvm.org/docs/WritingAnLLVMPass.html#LoopPass>) is misleading or > incorrect (or if I'm just missing something.) The documentation states: > > "The doFinalization method ... is called when the pass framework has > finished > calling runOnLoop > <http://llvm.org/docs/__WritingAnLLVMPass.html#__runOnLoop > <http://llvm.org/docs/WritingAnLLVMPass.html#runOnLoop>> for > > every loop in the program being compiled." > > From this, I understood that for a single instance of a...

Hi I have the followed code for which I am writing a loop pass. int main() { int i = 0; for (i=0; i<20; i++) { printf ("hello world %d\n", i); } return 0; } In the function runOnLoop, I have the following instruction PHINode *indv = NULL; indv = L->getCanonicalInductionVariable(); However, when I check indv is always set to NULL. Since the code has a canonical induction variable, I was expecting the endv to start pointing to the induction var phi node. Please let me what I...

...ct which needs loop-attached metadata similar to the parallel metadata discussed in here. Currently I'm trying to make the metadata get through the optimisation passes unscathed. For loop passes, I've solved that by modifying the LPPassManager to check for metadata presence before and after runOnLoop, and restoring it if necessary (assuming the loop didn't get removed and the metadata didn't get moved). However, the CFG simplifier pass also can optimize the loop latch and attached metadata away. As this is a function pass, I don't have access to the loop-specific functionality such...

Hi, I am using Loop pass ( runOnLoop() function ) In this function I want to access all the Loops in all the functions in the current Module the LoopInfo Pass only gives the Loops in the function where the current Loop resides :-( Is there any other Pass by which I can access the Loops in others functions as well ? ...... Regards,...

...-> getLoopFor(bb): >> >> >> Two steps: >> 1) In your pass's getAnalysisUsage(), request LoopInfo. >> >> virtual void getAnalysisUsage(AnalysisUsage &AU) const { >> AU.addRequired<LoopInfo>(); >> } >> >> 2) In your runOnLoop() method get the analysis. >> >> LoopPass *LI = &getAnalysis<LoopInfo>(); >> >> Now, you can do LI->getLoopFor(...) >> >> >> To add to what Devang has said, you may find the doxygen docs on LoopInfo >> helpful: >> >> http://...

...nably simple way to identify loop(s) within a function under FunctionPass? (Any existing LLVM code that I can take a look?) OR Is it possible to intermix different types of LLVM Passes? E.g. class MyPass: public FunctionPass, public LoopPass{ ... public: bool runOnFunction(...); bool runOnLoop(...); ... }; Does this make any sense? I haven't seem to find any existing LLVM pass under such uses. Thank you very much Chuck -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20110106/b003904a/atta...