Taking advantage of GHC's SIMD support from within DPH is in the works. I'd love to have some good target applications to help drive the work. An application with DPH, plain C, C+SSE intrinsics, and, say, OpenCL versions would be even better, but that's probably wishful thinking. Dmitry, is your thesis about gas dynamics (I suspect it is) and you just want to try using Haskell, or is it about using Haskell to implement scientific code? Geoff On 03/21/2012 14:12, Ryan Newton wrote:

The big topic would seem to be using Geoff Mainland's new simd-ghc support.

http://hackage.haskell.org/trac/ghc/wiki/SIMD

Geoff, is that already under way or is it too big a project for a GSOC?

Dmitry, do you already have GHC-internals hacking experience?

-Ryan

On Tue, Mar 20, 2012 at 8:52 PM, Dmitry Dzhus mailto:dima@dzhus.org> wrote:

How are you gentlemen.

I'm interested about what additions to DPH would really benefit the community.

The only ticket on DPH in GSoC Trac seems a bit stale: http://hackage.haskell.org/trac/summer-of-code/ticket/1537. Any new tips?

My field is currently molecular simulation. (I'm doing a Direct Simulation Monte-Carlo implementation for rarefied gas dynamics with application to space satellites&stuffs aerodynamics in Haskell for my Master's thesis, I believe using DPH as well would be my final goal as molecular simulations vectorize nicely. Simulation domain is partitioned into cells and stochastic collision sampling (as in DSMC) may be performed in every cell in parallel.)

21.03.2012, 18:12, "Ryan Newton" :

The big topic would seem to be using Geoff Mainland's new simd-ghc support.

    http://hackage.haskell.org/trac/ghc/wiki/SIMD

Geoff, is that already under way or is it too big a project for a GSOC?

Dmitry, do you already have GHC-internals hacking experience?

That's interesting. I have experience with SIMD (more exactly, SSE) itself from Parallel Programming course I had last year (which included writing SSE implementations of common stuff like DFT). As for GHC hacking, no, but learning in this would be a really interesting goal for me. Perhaps a GHC-related assignment from possible mentor could be a prerequisite for any project proposal from my side, to see how quickly I can dive into certain details of GHC implementation.