Hi Sahil --

Sorry for the delayed response.  A good starting point for Polly-based
study of Chapel might be our ports of the PRK Stencil benchmark which
lives in this directory:


I'm suggesting this one because my understanding is that stencils are ripe for polyhedral optimization as in systems like Polly, because it's fairly
simple, and because it's a community benchmark (see more about the PRK
suite here:



Best of luck with your GSoC application,

On Wed, 28 Mar 2018, SAHIL GIRISH YERAWAR via Chapel-developers wrote:

Hi all

I am Sahil Yerawar, 3rd Year Btech Student at IIT-Hyderabad. I have
submitted the proposal "Compiling Chapel with LLVM/Polly" to Polly Labs for
GSoC 2018 and wish to initiate the work soon enough.

I have been spending time learning about Polly and how it works on LLVM IR
and will continue the process for some more time.

In addition to this pathway, I would also like to parallelly explore the
kernels i.e specific code snippets which are commonly used in the area of
High-Performance Computing. I believe that targeting these kernels and
enabling them to be recognized by Polly would be a more structured way to
approach this problem.

I would like to know from the Chapel-Dev Community about some instances of
these kernels in order to initiate a study of some kernels beforehand.
Feedback and suggestions are welcome.

Thanking you
Sahil Yerawar
Department of Computer Science
IIT Hyderabad
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
Chapel-developers mailing list

Reply via email to