A special request for GTX 470/480 owners Postby ihaque » Thu May 06, 2010 6:02 am Hi all, I know there are many of you out there eagerly waiting for the Fermi-compatible GPU client and core so you can do some folding. While you're waiting, I have a request for you. I'll be presenting a study on GPU reliability in a couple weeks, but I've got a severe lack of data on Fermi-based boards. Though I have a couple of boards in the lab to test on, that's really not a very big sample size. The proposition is simple: download MemtestG80 and run as much of it as you're willing over the next two weeks making sure to allow it to report results back to Stanford. Running more test iterations and a larger amount of memory is useful (to a point: if the amount of memory is too large, you'll start seeing some tests report > 1 billion errors; that's a timeout), but I'll take whatever data I can get. You won't get any Folding@home points for doing this, but you will have a grad student's gratitude - and the pride in knowing that your data will probably have the fastest turnaround time from submission to scientific presentation in FAH history... Imran MemtestG80 can be downloaded here: http://folding.stanford.edu/English/DownloadUtils PS - Feel free to pass this request on to your folding teammates Edit Doh...my apologies Hilbert...I should have linked to the original request. Here it is... http://foldingforum.org/viewtopic.php?f=52&t=14465
Maybe someone will fill me in here, because my knowledge is very limited. I thought the point of CUDA was to standardize programs for all Nvidia GPUs. Why then does Fermi require a new release?
where is this... ive looked on stanford site but cannot see 1.1 fermi not folded for some time and im still 25th ive got a new 480gtx and want to try it see what temps i get. thanks,
Here's the link (http://folding.stanford.edu/English/DownloadUtils) Do note that MemtestG80 1.1 is available for Windows and Linux. Just to check it, I ran it on my GTX 260 SOC and it ran successful.
