As of my checkin of June 1 2009, the following are scaled by #CPUs and/or #GPUs:
max_wus_to_send max_wus_in_progress daily_result_quota This makes sense for computational jobs, but not for non-CPU-intensive. Carl, if you upgrade QCN to current code you'll have this problem. The workaround is for non-CPU-intensive projects to use <max_ncpus>1</max_ncpus> It might be cleaner to have a <dont_scale_njobs> flag instead. -- David Carl Christensen wrote: > in the server config.xml on qcn I have: > > <max_wus_in_progress> > 1 > </max_wus_in_progress> > > that enforces just one job for multiprocessor machines > > > _______________________________________________ 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.
