Bug in server affecting older BOINC clients with NVIDIA GPUs.


log in

Advanced search

Message boards : News : Bug in server affecting older BOINC clients with NVIDIA GPUs.

Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · Next
Author Message
Paulie
Send message
Joined: 10 Nov 11
Posts: 6
Credit: 1,473,296
RAC: 2,218
United States
Message 1278486 - Posted: 31 Aug 2012, 22:05:32 UTC - in response to Message 1278443.

Thanks.

I saw a dozen or so tasks running. I have my preferences set at 75%.

Since I just installed this new graphics card, must be the new computing power ...

Paulie
Send message
Joined: 10 Nov 11
Posts: 6
Credit: 1,473,296
RAC: 2,218
United States
Message 1279144 - Posted: 2 Sep 2012, 7:31:39 UTC - in response to Message 1278486.

Just to verify, running about 14 SETIs at one time is OK ?

rob smithProject donor
Volunteer tester
Send message
Joined: 7 Mar 03
Posts: 8403
Credit: 56,917,574
RAC: 76,647
United Kingdom
Message 1279150 - Posted: 2 Sep 2012, 7:40:57 UTC - in response to Message 1275574.

Keith,
Are other projects running at "high priority"
I ask this because some other projects send out work with very short deadlines (compared to S@H) that they effectively "bully" S@H off the scene for a bit. It is a pain somewhere below the waistline....
____________
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Profile Sunny129
Avatar
Send message
Joined: 7 Nov 00
Posts: 190
Credit: 3,163,755
RAC: 0
United States
Message 1279164 - Posted: 2 Sep 2012, 8:27:56 UTC - in response to Message 1279144.

Just to verify, running about 14 SETIs at one time is OK ?

no...definitely not on a single GTX 550 Ti. w/ 2 Multibeam tasks running in parallel on one of my GTX 560 Ti's, VRAM consumption sometimes maxes out above 700MB. some quick math tells us that trying to run 3 tasks in parallel will almost certainly require more than the 1GB of VRAM on the GPU. some more quick math tells us that 14 Multibeam tasks would require in the neighborhood of ~5GB of VRAM. whenever you over-allocate either your GPU's core and/or memory resources, you sacrifice crunching efficiency and increase run times.
____________

musicplayer
Send message
Joined: 17 May 10
Posts: 1431
Credit: 687,186
RAC: 0
Message 1279167 - Posted: 2 Sep 2012, 8:35:09 UTC
Last modified: 2 Sep 2012, 8:48:51 UTC

Grrr! Why not ask me - PrimeGrid, of course.

Long tasks - incredibly short deadlines.

PrimeGrid tasks are running at high priority according to BOINC Manager. Seti@home gets second in the queue just because of these short deadlines.

But please do not try to adjust the priorities of tasks manually using Windows Task Manager. I did such a thing earlier on and had to restart my computer possibly because as a result of receiving a "blue screen" error.

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4100
Credit: 33,069,358
RAC: 8,872
United Kingdom
Message 1279177 - Posted: 2 Sep 2012, 9:02:57 UTC - in response to Message 1279144.

Just to verify, running about 14 SETIs at one time is OK ?

Your host should run 3 tasks at one time, two CPU tasks with the 6.03 MB CPU app or the 6.01 AP CPU app, and one task on the GTX550Ti using eithier the 6.10 cuda_fermi app or one task using the 6.04 OpenCL AP app,

None of your completed Cuda_Fermi tasks are showing any signs of running multiple tasks at once, ie extended runtimes, or CPU fallback, or erroring out,

Are you sure you're not confusing 'Running' tasks with 'Ready to Start' or 'Waiting to Run' tasks.

Claggy

Profile Sunny129
Avatar
Send message
Joined: 7 Nov 00
Posts: 190
Credit: 3,163,755
RAC: 0
United States
Message 1279298 - Posted: 2 Sep 2012, 16:27:46 UTC
Last modified: 2 Sep 2012, 16:28:35 UTC

thanks for the insight Claggy...i looked at his platform details, but it didn't dawn on me to check his actual tasks...so i kind of assumed he was trying to run 14 tasks in parallel on his GTX 550 Ti LOL...my mistake. i see now that he's only running 2 at a time on his dual core CPU and one at a time on the GPU. that being said, don't you think Paulie might stand to bring down his GPU task run times by trying to run 2 GPU tasks in parallel? his GPU has enough VRAM for it...i think its just a matter of whether or not the GPU core itself becomes overloaded w/ only 2 tasks running. i think it might be worth a shot...after all, i get reduced run times when running 2 GPU tasks in parallel vs only 1 task at a time on my GTX 560 Ti's.
____________

Paulie
Send message
Joined: 10 Nov 11
Posts: 6
Credit: 1,473,296
RAC: 2,218
United States
Message 1279327 - Posted: 2 Sep 2012, 18:08:15 UTC - in response to Message 1279298.

How do I do this?

Paulie
Send message
Joined: 10 Nov 11
Posts: 6
Credit: 1,473,296
RAC: 2,218
United States
Message 1279331 - Posted: 2 Sep 2012, 18:19:12 UTC - in response to Message 1279327.

I clicked on run while computer in use and see I'm only running two SETIs, the others are apparently on standby ...

Profile Sunny129
Avatar
Send message
Joined: 7 Nov 00
Posts: 190
Credit: 3,163,755
RAC: 0
United States
Message 1279463 - Posted: 3 Sep 2012, 1:45:15 UTC - in response to Message 1279327.

How do I do this?

I clicked on run while computer in use and see I'm only running two SETIs, the others are apparently on standby ...

that's b/c you're running the stock application, which only allows for 1 GPU task to run at a time. in order to run more than one task on a single GPU simultaneously, you must employ what is known as an app_info.xml file in your SETI@Home project folder. there is a line of code in this file that you must manipulate in order to change the number of tasks you'd like your GPU to crunch simultaneously. but that's a topic for an entirely different thread. search the forums, as there's plenty of info to be found on app_info.xml files, and how to employ and manipulate them.
____________

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2694
Credit: 6,093,820
RAC: 4,644
Bulgaria
Message 1279957 - Posted: 4 Sep 2012, 11:55:58 UTC - in response to Message 1277877.
Last modified: 4 Sep 2012, 12:08:51 UTC

ID: 6017961

Are you sure?
This computer shows 'Last contact' with servers on 29 Jul 2012
And 'ATI Radeon HD 2300/2400/3200' GPU and not nVidia

And it have only 3 CPU tasks 'SETI@home Enhanced v6.03' still unreported:
http://setiathome.berkeley.edu/results.php?hostid=6017961


1 yr old. PSU: 12v/14A, DC pwr 250W

These 2 values seem too small (or at the edge) to me to feed:
AMD Athlon(tm) II X4 645
ATI Radeon HD 2300/2400/3200
HDD
DVD ...

... or the PSU (especially the capacitors) is aging

One year is enough time for accumulated dust to block the CPU/GPU heatsink fins (under the fan) and to cause CPU/GPU overheat

And again: you do not appear to have (had) SETI GPU tasks (ever) so GPU can't be heated by SETI
http://setiathome.berkeley.edu/host_app_versions.php?hostid=6017961


I'll bury it in the backyard and grieve for my lost data.

Your data (e.g. photos, documents, ...) is on the HDD and is very probably untouched and OK
Put this HDD in another computer and you will find your data probably in D:\Users\<your-name>\... (find here Desktop or Documents)


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Sean Farrell
Send message
Joined: 17 Jun 12
Posts: 2
Credit: 17,251
RAC: 0
Canada
Message 1280001 - Posted: 4 Sep 2012, 15:34:50 UTC - in response to Message 1275017.

I'm not blaming BOINC, just an observation....besides I'm sure i gave up the right to blame when i installed it..my 2 cents

Profile ex_brit
Avatar
Send message
Joined: 14 Feb 04
Posts: 182
Credit: 427,492
RAC: 25
Canada
Message 1280202 - Posted: 5 Sep 2012, 10:00:17 UTC - in response to Message 1271871.

We've identified a bug in the current BOINC server that is online at SETI@home. With older BOINC clients this bug results in running multiple SETI@home GPU applications simultaneously on a single GPU.

While we debug and fix the problem we've suspended distribution of NVIDIA work. We hope that everything will be back to normal some time tomorrow.


I have the latest BOINC version and an Nvidia GTX 295. I'm getting an awful lot of Computation Errors with Seti work units.
Is this why?

____________
Peter.
Toronto, Canada

Josef W. SegurProject donor
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4247
Credit: 1,048,692
RAC: 265
United States
Message 1280259 - Posted: 5 Sep 2012, 14:02:10 UTC - in response to Message 1280202.

We've identified a bug in the current BOINC server that is online at SETI@home. With older BOINC clients this bug results in running multiple SETI@home GPU applications simultaneously on a single GPU.

While we debug and fix the problem we've suspended distribution of NVIDIA work. We hope that everything will be back to normal some time tomorrow.


I have the latest BOINC version and an Nvidia GTX 295. I'm getting an awful lot of Computation Errors with Seti work units.
Is this why?

No, the bug (which was fixed Aug 17) only affected those running older BOINC clients. Your 7.0.28 was immune.

The two errors showing in your task list show "197 (0xc5) EXIT_TIME_LIMIT_EXCEEDED" errors after over 4 hours of run time, and both are normal tasks which should have completed in less than half an hour on a GTX 295.
Joe

Profile ex_brit
Avatar
Send message
Joined: 14 Feb 04
Posts: 182
Credit: 427,492
RAC: 25
Canada
Message 1280264 - Posted: 5 Sep 2012, 14:10:01 UTC - in response to Message 1280259.

Hmm, that's interesting. I use TThrottle to throttle back both the GPU and CPU should certain temperatures be exceeded. Do you think that might be something nto do with that?
____________
Peter.
Toronto, Canada

wilfredo
Send message
Joined: 16 Jun 99
Posts: 1
Credit: 582,904
RAC: 656
Puerto Rico
Message 1280850 - Posted: 7 Sep 2012, 3:29:13 UTC - in response to Message 1271877.

what?

Josef W. SegurProject donor
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4247
Credit: 1,048,692
RAC: 265
United States
Message 1280888 - Posted: 7 Sep 2012, 5:55:56 UTC - in response to Message 1280264.

Hmm, that's interesting. I use TThrottle to throttle back both the GPU and CPU should certain temperatures be exceeded. Do you think that might be something nto do with that?

The basic math of the elapsed time cutoff is 10x the estimated runtime. I doubt TThrottle is reducing processing rate that much unless you have a serious overheating problem. There are hints in the stderr details that those tasks were only about 1/3 done when they were killed, too. If that's right the tasks were slowed by a factor around 30x.
Joe

Profile ex_brit
Avatar
Send message
Joined: 14 Feb 04
Posts: 182
Credit: 427,492
RAC: 25
Canada
Message 1280940 - Posted: 7 Sep 2012, 10:09:22 UTC - in response to Message 1280888.

Hmm, that's interesting. I use TThrottle to throttle back both the GPU and CPU should certain temperatures be exceeded. Do you think that might be something nto do with that?

The basic math of the elapsed time cutoff is 10x the estimated runtime. I doubt TThrottle is reducing processing rate that much unless you have a serious overheating problem. There are hints in the stderr details that those tasks were only about 1/3 done when they were killed, too. If that's right the tasks were slowed by a factor around 30x.
Joe


Thanks - I'll keep an eye on it. So far so good since then.
____________
Peter.
Toronto, Canada

JimSEan
Send message
Joined: 19 Mar 01
Posts: 2
Credit: 475,734
RAC: 403
United States
Message 1281011 - Posted: 7 Sep 2012, 14:20:28 UTC - in response to Message 1271871.

What do I need to do now?

Jim Dugan
____________

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4100
Credit: 33,069,358
RAC: 8,872
United Kingdom
Message 1281142 - Posted: 7 Sep 2012, 18:55:27 UTC - in response to Message 1281011.

What do I need to do now?

Jim Dugan

Nothing, as you're not running an older Boinc client, and you don't have a Nvidia GPU!

[Off Topic]
If you upgrade your drivers for your HD5x00 you should be able to get that crunching.
[/Off Topic]

Claggy

Previous · 1 . . . 3 · 4 · 5 · 6 · 7 · Next

Message boards : News : Bug in server affecting older BOINC clients with NVIDIA GPUs.

Copyright © 2014 University of California