If it's running cooler than normal, then it's not working. These aren't short tasks, they're errors. Unless you track the problem down, you'll be wasting your time and electricity.

Edit - here is one of yours, against a 'normal' wingmate:

1386684476 5082452 10 Oct 2009 17:54:15 UTC 11 Oct 2009 11:15:51 UTC Completed, validation inconclusive 5,597.38 5,147.71 31.92 pending SETI@home Enhanced v6.03
1402028721 2612987 24 Oct 2009 12:43:42 UTC 24 Oct 2009 15:40:09 UTC Completed, validation inconclusive 50.38 48.63 0.01 pending SETI@home Enhanced v6.08 (cuda)

Yours is the one which ran for 50.38 seconds, 48.63 of them on the CPU. 2 seconds of GPU processing doesn't generate much heat!

NB: We often say on these boards that "Informational message -9 result_overflow" is not an error, but I'm afraid it is if your wingmate doen't see the same noise in the data.

Good thought, but I had already checked that...actually running cooler than normal. It's funny that BOINC is requesting as many as 40 new tasks at a time.

Anyway, will try re-boot and BOINC restarts and report back.

The request is in seconds, not tasks.

What are the Long Term and Short Term Debts of your projects? These can be found in the client_state.xml file in the BOINC data directory. The BOINC Data directory can be found in the first few messages after a restart of BOINC.

---

BOINC WIKI

Good thought, but I had already checked that...actually running cooler than normal. It's funny that BOINC is requesting as many as 40 new tasks at a time.

Anyway, will try re-boot and BOINC restarts and report back.

---

It does look like it is on your end. Try shutting down and restarting BOINC completely. (Advanced tab, shut down connected client, exit BOINC) If that doesn't work then try a reboot of your computer. Also, if you can, check the temps on your GPU. It could be running hot or it may be going out on you. If the temps are high try blowing out the dust bunnies. If none of that helps then I'm afraid it's time for a new GPU.

---

PROUD MEMBER OF Team Starfire World BOINC

Have you tried restarting your machine? your pending ones seems to suggest they are all going to -9 result overflows, whereas your wingman in atleast one of them managed to get a more likely result. A restart might help clear it so they process correctly

---

Suddenly my queue has about 100 seti gpu work units. Each takes about 45 seconds to execute. Boinc uploads them, reports them done in batches of 20 or 30, then gets dozens more. The problem is that GPUgrid never runs, and seeing the queue of seti cuda work in front of it, doesn't even ask for more work. But Seti keeps on sending these tiny wu's by the hundreds.

Up until now the Seti GPU work units have taken about 30 minutes to run, and there are generally at the most 2 or 3 in the work queue.

Oh, and also, when I looked up my pending credits, these little things are only worth 0.01 credit. So that works out to maybe 8 credits/hour. Crunching GPUgrid on my card yields maybe 500 credits/hour.

So I've been a SETI supporter for 10 years, over a million credits to my name....but, hey, I don't work for free!

Anyone know what's going on?

---