i started to have that problem .
the way i got it to start to work was because it was only a cpu job, i just suspended the task only for a few days and let all the others go. i then un-supended it and i now have no problem.
that happen on my 7.025, 7.0.26 and my 7.0.27 boxes

Set ALL your projects to "no new tasks".
Let them all complete and report.
Then revert to version 6 and download new work.
A technical code issue in version 7 causes this problem and will be fixed in a newer upcoming version in the future.

Yeah, no. If milkyway is the program that's malfunctioning, then I won't be not doing other projects for its sake. Instead, rather, I'll just stop crunching for Milkyway.

The problem is they went to a new system of determining which Project should run at what internal priority and MW is low on the list in the beginning. If you just leave it alone it will work itself out, or you can upgrade to Boinc version 7.0.27.

like i said it was happening even on my 7.0.27 box. and the only way i got it to stop was to do a task suspend on the work unit for a few days until restart of the task.

Hi again, since I last posted on this thread on May 3rd after I had upgraded to 7.0.27 a strange thing has happened. Although the high priority tag has disappeared with 7.0.27 the fault still appears with MW@H kicking SETI of my cpu.

I had thought that it would all sort itself out but 14 days later it is still happening and I am no nearer 50/50 SETI / MW@H. This is all complicated by the fact that SETI tasks have about 30+ days deadline and MW@H about 12 days but it should manage an even distribution.

Further complications this week with SSL at Berkeley having a power outage so I cannot contact the BOINC forums and my completed SETI tasks wont upload as both are based there.

I am going to leave things running and see where I get to in another 14 days. The good news for MW@H is "what you lose on the roundabouts you gain on the swings" so MW@H will get my undivided attention and the 50/50 problems will just have to wait until things settle down a bit.

John

Hi again, since I last posted on this thread on May 3rd after I had upgraded to 7.0.27 a strange thing has happened. Although the high priority tag has disappeared with 7.0.27 the fault still appears with MW@H kicking SETI of my cpu.

I had thought that it would all sort itself out but 14 days later it is still happening and I am no nearer 50/50 SETI / MW@H. This is all complicated by the fact that SETI tasks have about 30+ days deadline and MW@H about 12 days but it should manage an even distribution.

Further complications this week with SSL at Berkeley having a power outage so I cannot contact the BOINC forums and my completed SETI tasks wont upload as both are based there.

I am going to leave things running and see where I get to in another 14 days. The good news for MW@H is "what you lose on the roundabouts you gain on the swings" so MW@H will get my undivided attention and the 50/50 problems will just have to wait until things settle down a bit.

John

One thing you should be able to do is to do ONLY cpu crunching at Seti and ONLY gpu crunching at MW. That way they won't collide and then after say 6 months you could switch. This would require you to do a change on the website of each project saying you do NOT want a particular kind of unit from them.

Thanks Mikey but my wee NVidia GPU is not up to crunching.

John

Thanks Mikey but my wee NVidia GPU is not up to crunching.

John

Ahh sorry, I saw this and thought gpu:
212836827 168075260 19 May 2012 | 6:06:57 UTC 19 May 2012 | 8:53:03 UTC Completed and validated 129.77 230.04 1.87 MilkyWay@Home N-Body Simulation v0.80 (mt)
212828514 168068544 19 May 2012 | 5:52:33 UTC 19 May 2012 | 8:53:03 UTC Completed and validated 8,634.43 8,015.75 159.86 MilkyWay@Home v1.00

Sorry!

ps have you tried your gpu on Collatz? They take the most bare bones ones and make them work.

MW@H is still running SETI off my gpu. I only have two cores so I set it at 50/50 but it is not sorting itself out. When MW@H was running at high priority on 7.0.25 I thought that by upgrading to 7.0.27 that would sort the problem but it does not seem to be happening. I upgraded to 7.0.27 on the 3rd May and 17 days later I still have to intervene by suspending MW@H tasks for a while to allow SETI some time. Both projects used to run one on each core and I would like to get back to that.
I have to leave my pc unattended for a few days and if things are still going on like this then I will return to no SETI calculated.

Is there some way that I can intervene with BOINC and sort this out?

Thanks
John

Both projects used to run one on each core and I would like to get back to that.
John

That is NOT how Boinc is designed to run. Boinc runs on a debt based system, where when Seti is down Boinc figures that out and crunches MW, in your case, instead and 'owes' Seti time. MOST people see Boinc running one project than the other on their machines using all their cpu's doing it, swapping back and forth between projects as work and debts allow.

Exactly my point Mikey. Even after the 7.0.25 glitch has been "fixed" with 7.0.27 MW@H is running at high priority and has been for 18 days. If it were working correctly the "debt" would be the other way as using 7.0.25 MW@H was running at high priority for weeks and any "debt" would be owed to SETI.

Does anybody have any ideas as it is annoying to have to suspend MW@H tasks to allow SETI to run.

John

have you expiremented changing the resource share of both projects? try leaving mw@h with resource share of 1 and seti at 100, or some other proportion like this.

most likely, boinc is interpreting mw@h as having a much shorter deadline and focusing on it, only running seti when the deadline is close.

Also, you could try shortening the 'switch between applications every xx minutes' setting under computing preferences.