running out of work--BRP4 consequence

Joined: 6 Dec 05
Posts: 2,821
Credit: 3,294,391,991
RAC: 2,541,868
Topic 197261

I confess that I read this message over on the technical news section shortly after it was posted, but failed to notice that when it said the BRP4 work for CPUs had been stopped, that it meant I'd lost the CPU work source I had all my hosts set to. (I also run Perseus Arm GPU work--not affected in this)

The first symptom I noticed was that I seemed hardly ever to have tasks in the "ready to report" state. The second thing I noticed was that my hosts with the shortest queues started downloading and doing SETI work.

The explanations are simple--I'm just posting this in case it might help someone else who reads here but did not spot the technical news item or digest the import to them who also was running a restricted to BRP4 configuration. One reason one might not notice is that the item is NOT called BRP4 on the Einstein preference page, but rather is called: Binary Radio Pulsar Search (Arecibo)

The absence of ready to report work was a consequence of BOINC frequently requesting CPU work for a queue increasingly far below set point.

And the "failover" to SETI was exactly as intended--as a host went completely dry of Einstein CPU work it requested and got SETI work (which I have set to zero work share with the intention of getting this behavior).

Anyway, I altered my Einstein preference to allow both CasA and Gamma-Ray work on all venues, and have already received some of each. I'll sort out which I prefer later.