No GPU work on one machine


log in

Advanced search

Message boards : Number crunching : No GPU work on one machine

Previous · 1 · 2
Author Message
LadyL
Volunteer tester
Avatar
Send message
Joined: 14 Sep 11
Posts: 1679
Credit: 5,230,097
RAC: 0
Message 1178280 - Posted: 15 Dec 2011, 12:48:01 UTC - in response to Message 1178272.

I stick mine right under the version number.

You will need to add those numbers to all 6 occurrences of the GPU part.


Thank you

I shall stop rescheduling and let the numbers stabilise and then give it a try.

Just before you add the <flops> entries, I suggest you use the rescheduler to protect against -177 errors for maximum elapsed time exceeded on cached tasks. The expert tab has a setting called something like "limit rsc_fpops_bound" which provides that protection when you do a reschedule, even if the other settings mean that no tasks are actually rescheduled. That's as much as I know, I hope someone with practical experience will provide a clearer explanation.

Those cached tasks will have scaled down rsc_fpops_est which is used to calculate estimated time, and rsc_fpops_bound which is used to calculate the time limit. The big increase in flops will make the times much shorter, but DCF will soon increase to fix the estimates. Unfortunately DCF isn't used for the limit.
Joe


Good info Joe, I just put <flops> onto my main cruncher, got -17's. Will give this a try. Or is there a way to edit fpops in the app_info?

Regards,

A


You'd have to increase every single rsc_fpops_bound entry by some 3 magnitudes - one per task. Using the rescheduler is by far the less errorprone option.

Kevin Olley
Send message
Joined: 3 Aug 99
Posts: 368
Credit: 35,177,995
RAC: 2,562
United Kingdom
Message 1178306 - Posted: 15 Dec 2011, 16:12:24 UTC - in response to Message 1177375.

I stick mine right under the version number.

You will need to add those numbers to all 6 occurrences of the GPU part.


Thank you

I shall stop rescheduling and let the numbers stabilise and then give it a try.





Seen running with flops for a couple of days, no problems, its all running good.



____________
Kevin


AndyJ
Avatar
Send message
Joined: 17 Aug 02
Posts: 248
Credit: 27,380,797
RAC: 0
United Kingdom
Message 1178323 - Posted: 15 Dec 2011, 18:03:30 UTC - in response to Message 1178280.

I stick mine right under the version number.

You will need to add those numbers to all 6 occurrences of the GPU part.


Thank you

I shall stop rescheduling and let the numbers stabilise and then give it a try.

Just before you add the <flops> entries, I suggest you use the rescheduler to protect against -177 errors for maximum elapsed time exceeded on cached tasks. The expert tab has a setting called something like "limit rsc_fpops_bound" which provides that protection when you do a reschedule, even if the other settings mean that no tasks are actually rescheduled. That's as much as I know, I hope someone with practical experience will provide a clearer explanation.

Those cached tasks will have scaled down rsc_fpops_est which is used to calculate estimated time, and rsc_fpops_bound which is used to calculate the time limit. The big increase in flops will make the times much shorter, but DCF will soon increase to fix the estimates. Unfortunately DCF isn't used for the limit.
Joe


Good info Joe, I just put <flops> onto my main cruncher, got -17's. Will give this a try. Or is there a way to edit fpops in the app_info?

Regards,

A


You'd have to increase every single rsc_fpops_bound entry by some 3 magnitudes - one per task. Using the rescheduler is by far the less errorprone option.


Thanks, seems to be working fine, will keep a close watch though.


Regards,

A
____________

Profile gcpeters
Avatar
Send message
Joined: 20 May 99
Posts: 67
Credit: 63,946,907
RAC: 29,828
United States
Message 1178336 - Posted: 15 Dec 2011, 20:07:27 UTC

I just noticed on my Nvidia GTX 460 box that it isn't getting squat either. No WUs at all...of any kind. All my other systems are chugging along just fine, getting new WUs...etc.

What is this new glitch this month?

AndyJ
Avatar
Send message
Joined: 17 Aug 02
Posts: 248
Credit: 27,380,797
RAC: 0
United Kingdom
Message 1178341 - Posted: 15 Dec 2011, 20:58:59 UTC - in response to Message 1178336.
Last modified: 15 Dec 2011, 21:14:15 UTC

I just noticed on my Nvidia GTX 460 box that it isn't getting squat either. No WUs at all...of any kind. All my other systems are chugging along just fine, getting new WUs...etc.

What is this new glitch this month?


Running two 460's here too. Nada, same as you.
No tasks available.

Regards,

A

Edit. Forget that, just got seven to download, not a lot, but things are beginning to move.

Regards

A
____________

Profile gcpeters
Avatar
Send message
Joined: 20 May 99
Posts: 67
Credit: 63,946,907
RAC: 29,828
United States
Message 1178348 - Posted: 15 Dec 2011, 21:34:23 UTC

Interesting. Mine just started working too. What an amazing coincidence! Not. Must have stirred up the goblins at Berkeley.

AndyJ
Avatar
Send message
Joined: 17 Aug 02
Posts: 248
Credit: 27,380,797
RAC: 0
United Kingdom
Message 1178368 - Posted: 15 Dec 2011, 22:00:51 UTC - in response to Message 1178348.
Last modified: 15 Dec 2011, 22:08:47 UTC

Yup, that was my super secret Goblin Stirring Berkely Software, GSBS, (tm) coming online.

Regards,

A

:-)
____________

Profile gcpeters
Avatar
Send message
Joined: 20 May 99
Posts: 67
Credit: 63,946,907
RAC: 29,828
United States
Message 1178522 - Posted: 16 Dec 2011, 16:18:05 UTC

I spoke too soon. I got a few WUs and then the faucet closed up tight again. It's been 16 hours since a new WU has been received. Why always this flaky behavior???

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3860
Credit: 107,032,847
RAC: 98,261
United States
Message 1178528 - Posted: 16 Dec 2011, 16:28:50 UTC - in response to Message 1178522.

I spoke too soon. I got a few WUs and then the faucet closed up tight again. It's been 16 hours since a new WU has been received. Why always this flaky behavior???

Very odd I have been constantly bouncing off of the limit with mine.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile gcpeters
Avatar
Send message
Joined: 20 May 99
Posts: 67
Credit: 63,946,907
RAC: 29,828
United States
Message 1178565 - Posted: 16 Dec 2011, 17:57:59 UTC

I think someone at SETI is just jealous that I'm running BOINC on 40 core servers...so they penalize me periodically. From the bizarro download behavior I regularly see, I can only conclude such a preposterous assumption.

msattler
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38322
Credit: 560,153,137
RAC: 652,782
United States
Message 1178570 - Posted: 16 Dec 2011, 18:05:41 UTC - in response to Message 1178565.

I think someone at SETI is just jealous that I'm running BOINC on 40 core servers...so they penalize me periodically. From the bizarro download behavior I regularly see, I can only conclude such a preposterous assumption.

No, Virginia, it is not your Santa Claus. LOL.
I see the same preposterous messages even though I only have 8 rigs online.
____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

Profile Alien Hunter, USA
Send message
Joined: 6 Oct 01
Posts: 35
Credit: 2,626,474
RAC: 2,261
United States
Message 1178871 - Posted: 17 Dec 2011, 15:59:10 UTC

Help my AMD doesnt crunch very much at all. and the turn around time is long.its merged with
my intel and has no problems.
____________

Profile Bob Bainbridge
Send message
Joined: 21 Mar 01
Posts: 54
Credit: 48,346,970
RAC: 57,295
United States
Message 1178937 - Posted: 17 Dec 2011, 20:25:12 UTC

My Intel 970 machine with GTX480 has been running for weeks now with little or no WU's. Hardly any CPU WU's have been downloaded. Currently I have 7 GPU WU's ready to run and I get a message "Not requesting new tasks" when I connect. My AMD amchine with a GTX570 is doing a little better and has been downloading enough WU's to keep busy. My RAC is down from a one-time 50k to 16k now and I've slipped from No.1 in my group. My slower laptop and AMD 4400X2 are chunking along with all the work they can handle. What gives?????


Bob B.
____________

Previous · 1 · 2

Message boards : Number crunching : No GPU work on one machine

Copyright © 2014 University of California