World Community Grid - View Thread

andreic, if the only project which you're crunching for is WCG, then, instead of manually 'pushing' those Beta WU's to the front, why not let BOINC do the work for you?

I've got my 'Switch between applications' setting set to 7200 minutes - which automatically forces all WU's with a shorter than normal deadline (and this includes repair WU's) to the front of the queue. Worked a treat for the four I've managed to snag biggrin

The minimum value to enter for the switch time depends on deadline and cache size. Found here in steps that a 1 day cache works with a 4000 minute setting and a 1.9 day cache with a 2800 minute will cause an HP state on the <= 4.0 days deadline jobs. It's the sum of the cache+the switch time+the safety factor the client has in-build that determines if BOINC goes into panic state for one or the other shorter-deadline task.

One would not push that HP forcing too hard as when you do e.g. CEP2 on a quad, and depending on mix, you could end up with one or more CEP2 tasks in "waiting to run, LAIM" state as just happened here... 1 C4CW + HCMD2 tasks suspended, 2 CEP2 running in normal state and 1 CEP2 and 1 C4CW in HP and 3 CEP2 concurrent is for this box is exponentially inefficient.

Just FYI