Set your queue sizes to 0. "maintain enough work for" and "maintain additional work for"
-----Original Message----- From: boinc_dev [mailto:[email protected]] On Behalf Of attila csaba marosi Sent: Wednesday, March 13, 2013 11:11 AM To: [email protected] Subject: [boinc_dev] fetch minimal work Hello, How is --fetch_minimal_work intended to work? I run my clients and first they will fetch indeed minimal work (1 sec), but next they will request 0 secs and then idle forever. Is this behavior known (or maybe intended) ? What I want to achieve is that the client should not "cache" any work just the one it runs (in case it is shut down it won't have to abort many -- I use report results immediately and abort jobs on exit as well). The wiki states "Fetch only 1 job per device (CPU, GPU)", so does this mean fetch 1 job only and never again or fetch jobs 1 at a time (per device) ? Thanks, Attila _______________________________________________ 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.
