I don't know, and I've not heard of anyone who has tried.... DCF is to correct between the estimated time for a task, and the actual time.
If a project has two different applications, one with a DCF around 0.6 and the other around 0.3, it seems to me that if you cut the estimate of the first application in half, or doubled the second, that DCF would converge on a reasonable value. A 10% variation would be no big deal, it's when they're off by orders of magnitude that you run into problems keeping the queue reasonably full, or running over deadlines. I know that isn't perfect, especially when you mix CPU and GPU, but it can be done right now. On 3/7/2010 10:41 AM, Eric Myers wrote: > > WCG also has multiple apps that probably have large differences > in duration. How do they deal with it? > > -Eric > > > > On Sun, 7 Mar 2010, Rytis Slatkevičius wrote: > >> I am going to raise DCF per app question once again. >> >> PrimeGrid currently has 13 different applications with varying task >> durations and optimization levels for different CPUs. Also available >> are GPU >> apps. This makes it very hard to correctly estimate the duration of >> tasks, >> and it throws BOINC duration estimate very much off when you run more >> than >> one application type. For example if you ran an app which was >> overestimated >> and had short tasks, and then switched to an app which has long tasks >> that >> are overestimated, it would make it seem that they will run for months, >> which leads to user complaints and lost CPU time because of WUs being >> cancelled, sometimes with a few days of progress. >> >> This has become a number one problem being discussed in the PG forums. >> Therefore I'm going to ask to once again reconsider implementing DCF >> per app >> instead of having it per project only. >> >> > > > > _______________________________________________ > boinc_dev mailing list > [email protected] > http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev > To unsubscribe, visit the above URL and > (near bottom of page) enter your email address. _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
