Boy, for me, this'll be cutting it real close - far too close for my own comfort...

sniffle-sniffle...

Re. projects that may benefit from GPU processing:

The hastening of the completion of the HCC project due to the introduction of GPU processing surprised me. Although the speedup provided by a GPU was as I expected, I thought that only a minority of members would have or acquire suitable GPUs so that the sheer numbers of CPUs crunching HCC would outweigh the contribution of the GPUs. WCG obviously now has a huge computing resource available in the form of GPUs.

It has been said many times here in the forum that none of the current projects can be ported to GPUs, so this resource will sit idle unless a new project that can use it comes to WCG, or members switch their GPUs to contribute to other distributed computing projects outside of WCG.

So OK, what do you do if you haven't got something? One option is to borrow/beg/steal someone else's! Existing GPU projects that would fit WCG's philanthropic goals are mostly simulating protein folding, eg POEM, GPUgrid and Stanford PF (non-BOINC). Some of the scientists working on those projects may be interested in running a project on WCG. WCG currently has one protein folding project: HPF2. It uses the Rosetta program, which can't benefit from running on a GPU. Perhaps the HPF scientists would be interested in performing some HPF calculations on GPUs using a different program that already exists. I don't know whether the results would be comparable, but there's a way to find out ...

Many thanks, Seippel.

Maybe others will learn from the HCC GPU experience. The important thing to me is that the research is being completed and I hope the results can be used to benefit cancer patients. Yesterday, I talked to an acquaintance who told me he has cancer. I know so many people who have died from cancer, are in remission or are being treated for cancer of many different types. Even if our WCG efforts help a little, the effort and expense will be worthwhile.

Once HCC completes, there are other medical research projects using GPU computing, notably GPUGrid - NVIDIA only (HIV,cancer, neurological disorders), Folding at Home - ATI and NVIDIA (my interest is in Alzheimer's) and POEM at home. POEM is no longer exclusively medical. If points/credits are your thing, GPUGrid and POEM deliver. Folding gives a low points score, however there is a plan to significantly increase the score in the near future.

There may be other projects out there....I will be looking as I have a few GPUs to employ...

Happy crunching folks!!

To be honest, the combined production of the GPU's surprised us as well. We had anticipated that with the launch of GPU we would double the daily output. instead it increased by a factor of 16-17 (before GPU we were completing ~60k/results per day with 1 image per job, we are now >500k/results per day with 2 images per job. This led to some great adventures in handling greater capacity (which we call a 'success' problem).