DPH and CUDA status

Jeff Heard jeff at renci.org
Wed Feb 3 12:34:39 EST 2010

IIRC, there has been woirk done by Manuel and his team.  I'm sure
he'll chime in on that.  One thing though, is that CUDA is being
supplanted by OpenCL in the next few years, and OpenCL can handle data
parallelism on multicore CPUs as well as GPUs with the same code.
It's a little more flexible overall than CUDA, and will be portable
across ATI/nVidia/Intell/AMD/Sony Cell in the end, and is well
supported on Linux, Mac, and Windows systems.

There are two Haskell bindings right now for OpenCL, mine which is
OpenCLRaw, and one that you can find IIRC on Google Code if you search
for Haskell and OpenCL in the same search string.

I would love to see a DPH backend in OpenCL, or failing that,
something like GPipe but for computational kernels instead of graphics
done in OpenCL.

-- Jeff

On Wed, Feb 3, 2010 at 12:28 PM, Felipe Lessa <felipe.lessa at gmail.com> wrote:
> Hello,
> Recently I tried to look for the status of Data Parallel Haskell
> with a CUDA backend.  I've found [1] which mentions [2] saying
> that this is difficult and work was being done.  That was almost
> two years ago.  Was any progress made since then or is the work
> stalled?
> About GSoC, I wonder if there is any part of the ticket that can
> be done in a summer's worth of time?
> Thanks, guys! :)
> [1] http://hackage.haskell.org/trac/summer-of-code/ticket/1537
> [2] http://article.gmane.org/gmane.comp.lang.haskell.cafe/37538
> --
> Felipe.
> _______________________________________________
> Glasgow-haskell-users mailing list
> Glasgow-haskell-users at haskell.org
> http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

More information about the Glasgow-haskell-users mailing list