[parsec-users] Porting PARSEC to GPUs (CUDA)

jcupitt at gmail.com jcupitt at gmail.com
Fri Feb 11 12:22:23 EST 2011

On 8 February 2011 19:48, Matt Sinclair <msinclair at wisc.edu> wrote:
>  I have done some work porting some of the PARSEC benchmarks to
> CUDA myself, and I'm interested to see if anyone else has done some
> work with different benchmarks that I could leverage?

We had a student who did some of vips in OpenGL a few years ago, but
it was a lot of work for a rather a small speed increase and I think
that's now been dropped.

The current plan is to use Orc (a run-time vector code generator) and
stay on the CPU. We get a useful x5 or so speedup over plain C for
things like convolution.


More information about the parsec-users mailing list