"I don't have time to touch the computer every 6 hours."

Can you explain this bit. Don't have to 'touch' anything, it's either the max In Progress allowed or the buffer size setting in days, which ever is smaller, the client requesting backfilling whenever there's a gap.

edit: To underscore, the 35 active core is an overall limit, not per science at WCG. While not being adverse to a higher number, there's greater reluctance [over caching and panic mode kick-in risks] due the variability of the run-times. This is a reason why WCG already locked the Duration Correction Factor to a 1.0 to prevent clients recalculating actual TTC's based on latest completed, the servers taking care of assigning a most recent -average- runtime on latest returned work [which has a lapse time of X].

Last adjustment was in February from 25 to 35: https://secure.worldcommunitygrid.org/forums/wcg/viewpostinthread?post=484358

Rob,
At the moment I'm trying to "babysit" my little crunch farm to catch enough FAAH resends to make it to 5 years. I didn't give up when the end was announced because I thought I could make it. The SOP for that used to be to queue up a day or 2 of other work and then reset preferences for FAAH. That doesn't work when a 35-WU limit is just 6 hours of OET work. (Someone I know didn't appreciate that limit either and found a work-around to queue up lots of FAAH before the end) That is the reason why I said I needed to "touch" the computer every 6 hours.

If, as you wrote in your reply, the server "knows" the runtime of a WU based on my recent returned work, then it is possible to limit my queue based on expected duration rather than an arbitrary cap of 35. 35 WUs could be anywhere between , which could vary between 6 hours and 1 month! If work is being sent with a 9-day deadline, then I should be able to queue up 9 days of work, right (though I never went above 5 days).

One way I could get around this is to *forgo* my exclusive relationship with WCG and crunch for another project as well. Is that what WCG wants me to do? I am certainly interpreting this decision as sending that message. I have always tried to play by WCG's rules and this is a real insult to me. I searched the forums and couldn't find any explanation from WCG or discussion (this is a community) of the merits first.

Cheers

I don't blame you, but one shouldn't have to reach into a bag of tricks to get serious work done. I would have tried something like that had I realized that I was so close to not making 5 years on FAAH. If I had been able to fill up my queue to more than 1 day of FAAH work, I might have gotten enough to finish out 5 years before the work steam dried up. I didn't notice this 35-day limit until that time. It hadn't caused me a problem until then. I had likely done "babysittin'", but the workunits weren't so small as to run into the limit.

Cheers

The way round it, as I'm sure you know, would be select FAAH and CEP2. CEP2 is there to keep the computer ticking over while asking for more FAAH WUs to fill the spare cache.

Patrick

Good idea Patrick. However I am quite done with CEP2. It caused havoc on my crunchers. I would only go back to the CEP2 project only if I have crunched 5 years for all other active projects.


Cheers