Yea credit is all over the place you can get 79 points or 980 points for doing the seemingly the same -wcgdpf 15 run.

But if you want further proof then look through your result list and look for tasks that has also been done by a wing-man, then 9 times out of ten he will get a completely different credit amount than you do.(many hundreds of points diffence usually) And thats's for doing the exact same work unit???.

Good evening,

So, the batches listed between 2225-3336 are known bad batches that were given to us by the researchers. I am attempting to limit the reruns of those that are error state.

As for points, it is based 100% off of how your gpu runs those jobs within a work unit. We size the work units to be 20x larger than the CPU versions. If a work unit runs 100% for each job, then it'll get 20x the average credit that a CPU version gets. This is about 1600 points would be the max. What happens is that during the job runs they may only complete 80% of the estimated calculations needed. Thus we only grant 80% of the total 1600. This is why the values differ between batches and work units. During beta tests, we were seeing the average calculations to be close to 60% of the estimated calculations needed. This is good as it is the application running more efficiently.

Hope this helps explain the differences and the problems listed by nanoprobe.

Thanks,
-Uplinger