Compute Error

Message boards : Number crunching : Compute Error

To post messages, you must log in.

AuthorMessage
Profile AtheistRamblings

Send message
Joined: 17 Dec 05
Posts: 2
Credit: 151,245
RAC: 0
Message 62277 - Posted: 18 Jul 2009, 6:21:08 UTC
Last modified: 18 Jul 2009, 6:36:39 UTC

I've noticed that a number of workunits have tasks that have an outcome of "Client Error" and a client state of "Compute Error".

When I look at the task details I notice a number of "Can't acquire lockfile - exiting" messages.

Happens on two machines but not for all tasks. Both machines are using Boinc version 6.6.28. One machine is running Vista sp2 and the other Windows XP sp3. Boinc is running as a service on both machines. Both machines are dual-core (AMD).

Is there a reason/fix for this error?

Cheers,
Paul
ID: 62277 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5664
Credit: 5,724,885
RAC: 1,195
Message 62288 - Posted: 18 Jul 2009, 21:40:28 UTC - in response to Message 62277.  

I've noticed that a number of workunits have tasks that have an outcome of "Client Error" and a client state of "Compute Error".

When I look at the task details I notice a number of "Can't acquire lockfile - exiting" messages.

Happens on two machines but not for all tasks. Both machines are using Boinc version 6.6.28. One machine is running Vista sp2 and the other Windows XP sp3. Boinc is running as a service on both machines. Both machines are dual-core (AMD).

Is there a reason/fix for this error?

Cheers,
Paul



one thing first off. uninstall 6.6.x and install 6.4.7 for windows instead.
lockfiles have been discussed in the past.
you can find a solution here

unfortuantly i cant find anything further about lockfiles in the archives. we had a discussion about this a long time back, but i think that you can find these files in the directory of boinc - projects - slots and then look through the slot folders for boinc_lockfile. also some discussion about cpu not running at 100% causes these errors i think was part of it.
ID: 62288 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile AtheistRamblings

Send message
Joined: 17 Dec 05
Posts: 2
Credit: 151,245
RAC: 0
Message 62292 - Posted: 19 Jul 2009, 3:10:28 UTC - in response to Message 62288.  

Thanks for the info I'll look it up. The cpu not running at a 100% matches my case as I only give boinc 50% of CPU.

Cheers,
Paul
ID: 62292 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5664
Credit: 5,724,885
RAC: 1,195
Message 62293 - Posted: 19 Jul 2009, 7:22:26 UTC - in response to Message 62292.  

Thanks for the info I'll look it up. The cpu not running at a 100% matches my case as I only give boinc 50% of CPU.

Cheers,
Paul


If you want to reduce your "contribution" to rosetta, then go into your participants section here on the web page and lower the resource share in this section: Resource share and graphics Rosetta@home preferences.
You can still maintain 100% cpu usuage but the run time and the number or tasks you get from Rosetta will be lower.
ID: 62293 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Snags

Send message
Joined: 22 Feb 07
Posts: 198
Credit: 2,844,779
RAC: 895
Message 62295 - Posted: 19 Jul 2009, 9:29:20 UTC - in response to Message 62293.  

Thanks for the info I'll look it up. The cpu not running at a 100% matches my case as I only give boinc 50% of CPU.

Cheers,
Paul


If you want to reduce your "contribution" to rosetta, then go into your participants section here on the web page and lower the resource share in this section: Resource share and graphics Rosetta@home preferences.
You can still maintain 100% cpu usuage but the run time and the number or tasks you get from Rosetta will be lower.


Note that BOINC would still be running at 100% of (available) CPU time. The resource share preference allocates that 100% (or 50% or whatever you've set for "use at most x% of CPU time") among the different BOINC projects. If you are only attached to one project changing the resource share won't have any effect. If you are attached to multiple projects it would simply shift the resources from Rosetta to the other projects.

Many folks run BOINC with that setting at 100% without problems but if you already know that won't work for you perhaps you could experiment with letting BOINC use 100% of one CPU. Changing the "On multiprocessor systems, use at most x% of processors" to anything less than 100% will restrict use to only one of your two cores. I wonder if then increasing the CPU usage to 100% would prevent the lockfile problem from appearing while maintaining the overall responsiveness of your system to the level you require. Does anyone have any experience with this combination of settings they would be willing to share?

Snags
ID: 62295 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mikey
Avatar

Send message
Joined: 5 Jan 06
Posts: 1894
Credit: 8,791,381
RAC: 1,179
Message 62308 - Posted: 20 Jul 2009, 9:30:39 UTC - in response to Message 62295.  

Thanks for the info I'll look it up. The cpu not running at a 100% matches my case as I only give boinc 50% of CPU.

Cheers,
Paul


If you want to reduce your "contribution" to rosetta, then go into your participants section here on the web page and lower the resource share in this section: Resource share and graphics Rosetta@home preferences.
You can still maintain 100% cpu usuage but the run time and the number or tasks you get from Rosetta will be lower.


Note that BOINC would still be running at 100% of (available) CPU time. The resource share preference allocates that 100% (or 50% or whatever you've set for "use at most x% of CPU time") among the different BOINC projects. If you are only attached to one project changing the resource share won't have any effect. If you are attached to multiple projects it would simply shift the resources from Rosetta to the other projects.

Many folks run BOINC with that setting at 100% without problems but if you already know that won't work for you perhaps you could experiment with letting BOINC use 100% of one CPU. Changing the "On multiprocessor systems, use at most x% of processors" to anything less than 100% will restrict use to only one of your two cores. I wonder if then increasing the CPU usage to 100% would prevent the lockfile problem from appearing while maintaining the overall responsiveness of your system to the level you require. Does anyone have any experience with this combination of settings they would be willing to share?

Snags


I used to have my laptop running at 50% of available processors, so one only 1 core of the 2 would be used, and left the cpu usage at 100% and had no problems. I am back to 100% on both, for now, as my wife isn't noticing any slowdowns. I think I will go back to only 50% of the processors soon as it is a couple year old laptop and I don't want to stress it too much. I had an older single core Toshiba laptop and the hard drive died on it, twice! No it was not the cooling, it was the constant hard drive access, laptop hard drives aren't made for the access that desktop hard drives are. Although newer drives are getting better!
ID: 62308 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : Compute Error



©2024 University of Washington
https://www.bakerlab.org