similar to: [LLVMdev] Partial loop unrolling

Hi Rinaldini, In order to find information about loops inside a given function you should use something like "LoopInfo *LI = P->getAnalysis<LoopInfo>()", remembering to add "AU.addRequired<LoopInfo>();" to your getAnalysisUsage method. If the function you are interested to is not located in the module being compiled (if you created it as an auxiliary function,

Hi Cristianno, Thank you, it works :) with an extra cast:                                     Value *v ......                                     ConstantInt* RR = (ConstantInt *)v;                                     uint64_t VV = (RR->getValue()).getLimitedValue();                                     errs()<<"\nRR  "<<VV<<"\n";

Hi, Depending on what have run before your pass, the loop may have been unrolled or simplified if the computation inside the loop is too simple. Cheers, -- Arnaud de Grandmaison ________________________________________ From: llvmdev-bounces at cs.uiuc.edu [llvmdev-bounces at cs.uiuc.edu] On Behalf Of Cristianno Martins [cristiannomartins at gmail.com] Sent: Thursday, April 26, 2012 5:52 PM To:

Hello there, I'm writing a transformation pass for LLVM, and I hoped to use dce to clean up the resulting code after my pass. I just have some questions about LLVM's dce implementation. Well, my transformation is a function pass, and, after the changes are made, some instructions are not needed anymore. In order to easily get rid of those instructions, I'm setting all their uses to

Thank you for your answer! You are right, but I still don't know how to build such a list dynamically using LLVM API with user type inside (I don't know this type yet ofc). 2013/4/15 Cristianno Martins <cristiannomartins at gmail.com> > Hi, > > I guess you could use a template with a std::vector: you don't need to > define right away the type, if you create a

Alexandru Ionut Diaconescu wrote: > Hi Cristianno, > > Thank you, it works :) > > with an extra cast: > Value *v ...... > ConstantInt* RR = (ConstantInt *)v; Please use "cast<ConstantInt>(v)" instead. See http://llvm.org/docs/ProgrammersManual.html#the-isa-cast-and-dyn-cast-templates . Nick > uint64_t VV = (RR->getValue()).getLimitedValue(); >

Oops, that's true: I forgot about that XD sorry =) Glad I could help, -- Cristianno Martins PhD Student of Computer Science University of Campinas cmartins at ic.unicamp.br <cristiannomartins at hotmail.com> On Tue, Jun 25, 2013 at 4:38 AM, Alexandru Ionut Diaconescu < cyrusthevirus001x at yahoo.com> wrote: > Hi Cristianno, > > Thank you, it works :) > > with

Hi Cristianno, On 01/06/13 01:49, Cristianno Martins wrote: > Hello there, > > I'm writing a transformation pass for LLVM, and I hoped to use dce to clean up > the resulting code after my pass. I just have some questions about LLVM's dce > implementation. > > Well, my transformation is a function pass, and, after the changes are made, > some instructions are not

Hi, Sorry, but I'm not sure I understand what you're trying to do. I guess you want some kind of superclass of some set of classes that are defined in the llvm api(?). In that case, which are specifically the types you could have in this list you're trying to define? -- Cristianno Martins On Tuesday, 16 de April de 2013 at 04:22, B B wrote: > Thank you for your answer! >

Hello there, I'm trying to use a simple implementation of a queue (my own implementation of it, actually), and I'm trying to use the functions defined on my queue class inside some llvm-ir code. Unfortunately, the names of the functions are being messed up, then I created some function wrappers just to avoid having to deal with the C++ weird function renaming. So, to easily wrapper the

Hello again, First of all, thanks for all the answers =) they really helped a lot =D *Have you verified that some other pass is not adding the function declarations back in after your pass is executed (e.g., by using the -debug-pass=Executions argument to see what passes run after your pass)?* I considered that for a moment, but I realized that wouldn't be possible for two reasons: I

Hello ! This may be a trivial question, but I cannot get fields from a Value type. If my Value is i32 1, how can I store 1 in a different structure (an integer) ? For the type, I have getType(). For the value, I see no method in Value.h. I tried to cast to ConstantExpr and then to get operands, or cast to ConstantDataArray and then use getAsString(), but is not working. How should I get the

I am having a strange problem with loop unrolling. Attached is a small example that demonstrates what happens. There is a for-loop with a known trip count, and some control flow inside the loop. If the condition of the control flow only depends on the loop index and loop invariant variables, the loop is not unrolled. However, if the condition involves potentially loop variant variables, the loop

On Tue, May 14, 2013 at 11:42 AM, Cristianno Martins < cristiannomartins at gmail.com> wrote: > Hello there, > > I'm trying to use a simple implementation of a queue (my own > implementation of it, actually), and I'm trying to use the functions > defined on my queue class inside some llvm-ir code. Unfortunately, the > names of the functions are being messed up,

Cheng, Are you looking specifically for an analysis that can 'undo' the effects of loop unrolling, or do you want dependency analysis that can run on the loop prior to unrolling? For dependency analysis on loops (prior to unrolling) Preston and Sanjoy have been working on this, see: http://lists.cs.uiuc.edu/pipermail/llvmdev/2012-May/049769.html

Hi all, Attached herewith is a simple vectorized function with loops performing a simple shuffle. I want all loops (inner and outer) to be unrolled by 2 and as such used -unroll-count=2 The inner loops(with k as the induction variable and having constant trip counts) unroll fully, but the outer loop with (j) fails to unroll. The llvm code is also attached with inner loops fully unrolled. To

Thanks for the explanation. But I am a little confused with the following fact. Can't LLVM keep vectorizable_elements as a symbolic value and convert the loop to say; for(unsigned i = 0; i < vectorizable_elements ; i += 2){ //main loop } for(unsigned i=0 ; i < vectorizable_elements % 2; i++){ //fix up } Why does it have to reason about the range of vectorizable_elements? Even

Hi Alexandru, if you have a Value pointer named v, you could use the method v->getValue(). getLimitedValue(), which returns uint64_t, that can be casted as int. Hope I could help, -- Cristianno Martins PhD Student of Computer Science University of Campinas cmartins at ic.unicamp.br <cristiannomartins at hotmail.com> On Tue, Jun 25, 2013 at 4:05 AM, Alexandru Ionut Diaconescu <

If count > MAX_UINT-4 your loop loops indefinitely with an increment of 4, I think. On Thu, Oct 13, 2016 at 4:42 PM, Charith Mendis via llvm-dev < llvm-dev at lists.llvm.org> wrote: > So, I tried unrolling the following simple loop. > > int unroll(unsigned * a, unsigned * b, unsigned *c, unsigned count){ > > for(unsigned i=0; i<count; i++){ > > a[i] =

Hi Vivek, you can also include these lines below to your ~/.bash_profile: LLVM_BUILD="/Developer/llvm/build" # Path to your build directory alias new-clang="$LLVM_BUILD/bin/clang -Wno-expansion-to-defined -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/usr/include" alias new-clang++="$LLVM_BUILD/bin/clang++