search for: pouchet

...ild it seems to be covered here http://polly.llvm.org/get_started.html but any hints/comments/scripts would be very helpful - how to do the acceptance test and criteria for go/no-go is the make polly-test enough? or we need to run more extensive tests like PolyBench? http://www.cse.ohio-state.edu/~pouchet/software/polybench/ - polly release notes for the 3.2 release > > -bw > Pawel > On Nov 12, 2012, at 5:48 AM, Jack Howarth <howarth at bromo.med.uc.edu> wrote: > >> Tobi, >> It appears that a polly 3.2 branch wasn't created last night. Shouldn't ther...

Hi all, PolyBench (http://www.cse.ohio-state.edu/~pouchet/software/polybench/) is a well-known benchmark for polyhedral compiler. Since LLVM-Polly http://polly.llvm.org/) has provided a very good polyhedral optimizer for LLVM, could we add this benchmark to LLVM test-suite? I have attached the patch file to add PolyBench to LLVM test-suite. Best w...

...of PCP? could us just use the PCP language with annotation? > might be used to optimize code. This could give a first impression > what to expect from the polyhedral model in LLVM. > > There are also affords to establish an interchangeable polyhedral > format (scoplib - Louis-Noel Pouchet) and to generate a polyhedral > compilation package. These will allow to share/exchange optimizations > between different compilers and research tools. > > Furthermore an external interface will enable researchers to use LLVM > for their work on polyhedral optimizations. This migh...

...e http://polly.llvm.org/get_started.html > but any hints/comments/scripts would be very helpful > > - how to do the acceptance test and criteria for go/no-go > is the make polly-test enough? > or we need to run more extensive tests like PolyBench? > http://www.cse.ohio-state.edu/~pouchet/software/polybench/ > I apologize as I haven't seen previous mails about this topic. I think its fine to create the branch, and the polly team do their thing, but it will not be a part of the official LLVM 3.2 release process unless it is discussed in detail and decided by the LLVM Releas...

...ooPo (automatic parallelization), Pluto (optimization in general) or even Graphite might be used to optimize code. This could give a first impression what to expect from the polyhedral model in LLVM. There are also affords to establish an interchangeable polyhedral format (scoplib - Louis-Noel Pouchet) and to generate a polyhedral compilation package. These will allow to share/exchange optimizations between different compilers and research tools. Furthermore an external interface will enable researchers to use LLVM for their work on polyhedral optimizations. This might be useful as there is...

...tact are "Sjodin, Jan" <Jan.Sjodin at amd.com> Sebastian Pop <sebpop at gmail.com> Work on PCP has stalled because Jan has left for another group. In the hort term, Tobias is right that the best way to interface tools is to use the scoplib format and library from Louis-Noel Pouchet (PoCC). In any case, work on either PCP or scoplib will be beneficial. The community of polyhedral compilation experts is small, and its openly available output is already diminished by two proprietary development projects at IBM and Reservoir Labs. I strongly wish that any effort within LLVM...

It would help if you sent the IR you're giving to opt or at least a complete C function and your clang command line. ~Craig On Wed, Mar 14, 2018 at 3:05 PM, hameeza ahmed <hahmed2305 at gmail.com> wrote: > Hello, > > I m unable to vectorize following kernel by opt tool; > > for (i = 0; i < _PB_NI; i++) > for (j = 0; j < _PB_NJ; j++) > { >

Hello, I m unable to vectorize following kernel by opt tool; for (i = 0; i < _PB_NI; i++) for (j = 0; j < _PB_NJ; j++) { tmp[i][j] = 0; for (k = 0; k < _PB_NK; ++k) tmp[i][j] += alpha * A[i][k] * B[k][j]; } for (i = 0; i < _PB_NI; i++) for (j = 0; j < _PB_NL; j++) { D[i][j] *= beta; for (k = 0; k < _PB_NJ; ++k) D[i][j] +=

Hi All, Could anyone point me to some good benchmarking tools to measure the runtime performance of clang compiled C++ applications. Thanks ! - Jyoti -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20131003/3cc029f1/attachment.html>

...e dependence vectors for multi-dimensional arrays in nested loops, it needs to "delinearize" array element accesses to recover the subscripts in each dimension of the array. I believe that the current implementation of delinearization is based on this paper: http://web.cse.ohio-state.edu/~pouchet.2/doc/ics-article.15a.pdf. This paper describes how to delinearize the subscripts, and as a last step it requires certain conditions to be met in order to validate that the delinearized indexes are correct. The `tryDelinearize` function in `DependenceAnalysis.cpp` appears to be checking for cases...

Tobi, It appears that a polly 3.2 branch wasn't created last night. Shouldn't there be one now that polly is part of the llvm 3.2 release? Jack

Hi Pawel, Could you branch polly as well for the 3.2 release? We will need to contact the 'polly' guys to figure out how best to add it to the release (testing, etc.). -bw On Nov 12, 2012, at 5:48 AM, Jack Howarth <howarth at bromo.med.uc.edu> wrote: > Tobi, > It appears that a polly 3.2 branch wasn't created last night. Shouldn't there be > one now that polly

On 06/08/2013 11:26 PM, Star Tan wrote: > Hi all, > > > PolyBench (http://www.cse.ohio-state.edu/~pouchet/software/polybench/) is a well-known benchmark for polyhedral compiler. Since LLVM-Polly http://polly.llvm.org/) has provided a very good polyhedral optimizer for LLVM, could we add this benchmark to LLVM test-suite? > > > I have attached the patch file to add PolyBench to LLVM test-su...

...would like to have some feedback from the LLVM comunity to know if we need to change anything else. I've removed the problematic benchmarks from the Makefiles for now, so you should be able to run all of the benchmarks, and all of them should pass the tests. [1] http://www.cse.ohio-state.edu/~pouchet/software/polybench/ Cheers, Douglas -------------- next part -------------- A non-text attachment was scrubbed... Name: Polybench.zip Type: application/zip Size: 110061 bytes Desc: not available URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120530/861a40f1/attachment.zip>

...p at gmail.com> > > Work on PCP has stalled because Jan has left for another group. > Jan, what's the plan to integrate PCP in GCC and/or LLVM? > In the hort term, Tobias is right that the best way to interface tools is to > use the scoplib format and library from Louis-Noel Pouchet (PoCC). > > In any case, work on either PCP or scoplib will be beneficial. The community > of polyhedral compilation experts is small, and its openly available output > is already diminished by two proprietary development projects at IBM and > Reservoir Labs. I strongly wish that any...

....org/cpu/ . >> General Purpose CPU vendors use it to show performance improvements. >> > Waaay over tuned... > >> >> 3. There are also some other small benchmarks that could test compiler >> performance, like polybench (http://www.cse.ohio-state. >> edu/~pouchet/software/polybench/ <http://www.cse.ohio-state. >> edu/%7Epouchet/software/polybench/> ), which focus on evaluating the >> loop transformation of the compiler. >> >> I can't say with absolute certainty, but didn't these favor polyhedral > type loop optimiza...

...and better readability of the test cases. >> might be used to optimize code. This could give a first impression >> what to expect from the polyhedral model in LLVM. >> >> There are also affords to establish an interchangeable polyhedral >> format (scoplib - Louis-Noel Pouchet) and to generate a polyhedral >> compilation package. These will allow to share/exchange optimizations >> between different compilers and research tools. >> >> Furthermore an external interface will enable researchers to use LLVM >> for their work on polyhedral optimiz...

...personal perspective on the issue: Firstly I don't think TeX does pre-estimation of work, it's just the standard stepwise "refinement stuff". However, the idea itself isn't necessarily a bad one. In the academic community there's a set of people (John Cavazos, Lous-Noel Pouchet, others I've forgotten) doing work on estimating which sets of optimizations are most likely to produce a good speed up using estimation functions. (Note they're concerned primarily with outputted code speed, followed by the fact there are just too many combinations of optimizations to eval...

Adding LLVMdev, since this is intimately related to the optimization passes. > I think this is roughly because some function level optimisations are > worse than O(N) in the number of instructions. Please profile this and mail llvmdev regarding passes with significantly superlinear behavior (e.g. O(n^2)). My understanding is that these kinds of algorithmic problems are generally considered

...gt; multi-dimensional arrays in nested loops, it needs to "delinearize" array > element accesses to recover the subscripts in each dimension of the array. > I believe that the current implementation of delinearization is based on > this paper: >  http://web.cse.ohio-state.edu/~pouchet.2/doc/ics-article.15a.pdf > . > > This paper describes how to delinearize the subscripts, and as a last step > it requires certain conditions to be met in order to validate that the > delinearized indexes are correct. The `tryDelinearize` function in > `DependenceAnalysis.cpp` app...