Volunteer

[crb002]

Hey just got the message from PL01 notifying me about your project. I
would love to help out. Lots on my plate this weekend, but I should
get around to browsing the codebase by Monday and running some
benchmarks. Perhaps by then I can have something worth contributing.
Great work guys!

-Brew
http://blogofbrew.blogspot.com/2011/04/opencl-logjam.html?showComment=1307431595309#c9090956423922307340

[Viewon01]

Hi,

You're welcome. If you are interested to contribute, just tell me. As
you know there as some guys working full time in the corresponding
CUDA project, so it is just to tell that there is a lot of work and
interesting stuffs to do. Even more because we target more kind of
hardwares.

[crb002]

Yes, I think I am interested in contributing now that I have had a
chance to check out the code base.

My hardware is {Core2Duo(OSX), i7(OSX), AMD CPU(Linux/Windows),
9400m(OSX), RadeonHD4850(OSX), GTX460(Linux/windows)).

For now I think I will contribute some CPU radix sort code. The STL is
great, but I have yet to see an implementation that uses radix sort
under the hood. Not only will it make the benchmark comparisons more
honest, we will be able to have it default to an efficient CPU version
when the OpenCL device is a single core CPU.

Also, I will try to get some scripts going to pull versions from SVN,
and benchmark it on all 6 devices I have for regression tests.

-Brew

[kr...@polarlights.net]

Thanks for your contribution,

When we have start the project we have used the "stl sort" to compare the sort rate.
Our current CPU version is really faster and it is why we have remove the STL sort.

If you are interested by a pure CPU version, you can check the AMD samples.
You can also check Duane Merril web site, his method can also work on the CPU but
request some works. (Intel has implement it in one of their library I think).

Another interesting point will be to run it on the Fusion and Sandry Bridge APU.

Regards

Krys