Now if only my video cards would not keep dying on me.

---

Now if only my video cards would not keep dying on me.

Yep, I know that feelin'! [copyright an old ad for French parfum...]
Actually, people have been complaining that they can't get enough CPU WUs -- I seem to have the opposite problem: my faster machines keep asking for new CPU jobs, not GPU ones, and of course the scheduler says, "Nope, you've reached a limit!" E.g. the Supermicro has 2x4-core Xeons, not running hyperthreading, a C1060 and a GTX 460 so it should be getting an (8x50)+(2x400) limit, but it's struggling to get past 666 jobs in the queue because it's always asking for, and being denied, CPU jobs.
Answers on a postcard, please...

---

Answers on a postcard, please...

Just deselect 'Use CPU' in your project preferences until you're got lots of GPU tasks,

Claggy

Now if only my video cards would not keep dying on me.

Yep, I know that feelin'! [copyright an old ad for French parfum...]
Actually, people have been complaining that they can't get enough CPU WUs -- I seem to have the opposite problem: my faster machines keep asking for new CPU jobs, not GPU ones, and of course the scheduler says, "Nope, you've reached a limit!" E.g. the Supermicro has 2x4-core Xeons, not running hyperthreading, a C1060 and a GTX 460 so it should be getting an (8x50)+(2x400) limit, but it's struggling to get past 666 jobs in the queue because it's always asking for, and being denied, CPU jobs.
Answers on a postcard, please...

Maybe that's Yer problem Ivan... ;)

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

The answer to whats going on lies in technical news, latest post from mat about 3 hrs ago:-)

Regards,

---

Cliff,
Been there, Done that, Still no damm T shirt!

Now if only my video cards would not keep dying on me.

no ... you want them to die ... it's the excuse you need to get new ones ...

---

Bane is die! R.I.P.
You was a nice scheduler anyway. I'll miss you!

---

the only bane in my life is the fracking limits ... i would like to see these die as well ...

---

the only bane in my life is the fracking limits ... i would like to see these die as well ...

Not going to happen till the DCF problem is sorted.

---

Grant
Darwin NT

Now if only my video cards would not keep dying on me.

Yep, I know that feelin'! [copyright an old ad for French parfum...]
Actually, people have been complaining that they can't get enough CPU WUs -- I seem to have the opposite problem: my faster machines keep asking for new CPU jobs, not GPU ones, and of course the scheduler says, "Nope, you've reached a limit!" E.g. the Supermicro has 2x4-core Xeons, not running hyperthreading, a C1060 and a GTX 460 so it should be getting an (8x50)+(2x400) limit, but it's struggling to get past 666 jobs in the queue because it's always asking for, and being denied, CPU jobs.
Answers on a postcard, please...

Maybe that's Yer problem Ivan... ;)

Devilishly difficult, innit?

---

Well I'm empty and idle now. 10 days worth of APs have come and gone. Onward to waiting for the AP splitters to be updated and turned back on.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Well I'm empty and idle now. 10 days worth of APs have come and gone. Onward to waiting for the AP splitters to be updated and turned back on.

Get some MB to get your Johhny on, and then wait.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

I got a two seconds glimpse of something in one of my tasks.

Apparently BoincLogX said 13.50 for the gaussian score.

But checking SMV, I only get -2.613185 for the same score.

Task in question was 10ja12aa.24707.17823.8.10_132_0_0 .

Running by means of CUDA only.

But I may have been wrong.

Well its begining to take the biscuit, I have zero GPU tasks, load of CPU tasks and the server status page says tons of work available. However nearly every request for work gets the response no work available.

Does that mean that of all the tasks shown on the status page not one is a GPU task, that they are all CPU ones?

Over the past couple of hours on manual updates I've had 4 GPU tasks, if it were'nt for e@h my GPU's would be totally idle.

Would someone kindly give that sched server a tiny kick in the rear end and free the apparent logjam?
[edit]
Obliged to the prompt person at Berkeley, I just got 2 GPU tasks

---

Cliff,
Been there, Done that, Still no damm T shirt!

Ahh well at least I got 2WU last min or so,
but I'd have thought you would just love to have a lady jump all over you:-)

Cheers,

---

Cliff,
Been there, Done that, Still no damm T shirt!

I thought I'd put that into the FAQ. It's certainly been discussed often enough. ok, for the umpietst time...

a) There are no 'CPU' or 'GPU' tasks. they are all the same. some of them will be marked .vlar and those will not go out to CUDA GPU on account of bringing down most machines.

b) The feeder holds 100 tasks.it refills in regular intervals. the scheduler looks in the feeder if the type of task you are requesting (AP/MB) is available. if it isn't or if there is only vlar on a cuda GPU request or if the feeder is already empty you get 'no tasks available' if there were only vlar on GPU request you get 'no tasks sent'.

If you have trouble filling the cache for one resource but have plenty on the other, disable workfetch on the website for the resource that has plenty so onlyone of them will ask. Don't forget to enable again when you have enough.

P.S. You two can be thankful he found it funny. I didn't.

---

I'm not the Pope. I don't speak Ex Cathedra!

Something still seems wrong. RAC has been flat or falling on for the last four days, even on machines which didn't run out of work during the last outage

---

Thanks for the reminder, sorry if I upset you with my jest, it was not so intended.

Regards,

---

Cliff,
Been there, Done that, Still no damm T shirt!

wuid=945484895

Another symptom perhaps. This task appears to have received double credit. Is that possible?

---

AndrewM

wuid=945484895

Another symptom perhaps. This task appears to have received double credit. Is that possible?

WU true angle range is : 0.256601

That's a rare WU up near the peak for amount of processing which is needed. They take much longer to crunch and earn proportionally more credit, that one had a ~59 day deadline for instance.

                                                                   Joe