log in

Error in Computing

Message boards : Questions/Problems/Bugs : Error in Computing
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Kathryn Tombaugh-Weber

Send message
Joined: 12 Aug 11
Posts: 3
Credit: 6,884
RAC: 0
Message 758 - Posted: 19 Sep 2011, 14:03:07 UTC

The last several projects have all ended with errors in computing. Tried to change preferences, but am unable to save changes. What's going on?
ID: 758 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Greg
Project administrator

Send message
Joined: 26 Jun 08
Posts: 640
Credit: 433,775,088
RAC: 344,736
Message 759 - Posted: 20 Sep 2011, 21:27:19 UTC - in response to Message 758.  
Last modified: 20 Sep 2011, 21:27:29 UTC

You may want to switch to only receive 15e work units for a little while. We are getting closer to finishing up 2,1061-, the largest number we've attempted to date, and we are starting to push the sievers, especially the Windows versions, to their limits. An occasional computational error doesn't surprise me.
ID: 759 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Kathryn Tombaugh-Weber

Send message
Joined: 12 Aug 11
Posts: 3
Credit: 6,884
RAC: 0
Message 760 - Posted: 20 Sep 2011, 22:30:00 UTC - in response to Message 759.  

I actually just finished up a 15e task, and got no credit at all. I'm going to assume my computer is just not good enough for this project. Sorry.
ID: 760 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile ChertseyAl
Avatar

Send message
Joined: 11 Sep 09
Posts: 3
Credit: 123,354
RAC: 0
Message 762 - Posted: 7 Oct 2011, 17:42:28 UTC

I'm seeing the same problem. On one machine I'm getting quite a few errors of the same type as the OP. Half a dozen of my other machines seem OK.

Error WUs:

http://escatter11.fullerton.edu/nfs/result.php?resultid=15948413
http://escatter11.fullerton.edu/nfs/result.php?resultid=15947961
http://escatter11.fullerton.edu/nfs/result.php?resultid=15942169
http://escatter11.fullerton.edu/nfs/result.php?resultid=15942151

I'll go NNT on that one and keep an eye on the others. I don't recall having problems with this project in the past, but I've only just returned after a break for a while :)

Al.
ID: 762 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
David

Send message
Joined: 29 Dec 11
Posts: 1
Credit: 2,715
RAC: 0
Message 774 - Posted: 30 Dec 2011, 18:30:32 UTC

I know this is old news but I just have to say something.

I just signed in yesterday, 12/29. Downloaded many 16e & 15e workunits. My machine is an AMD FX-6100 6-core with 8 GB memory & a GTX460 video card. I only let BOINC use 5 cores at any one time. So 5 workunits started processing. The 1st 9 W/Us all erred out at between 1:53 & 1:56. So I started nosing around and found that they were consuming 70% of my 8 GB of memory and 100% of CPU time. The absolute highest memory usage I have ever seen. Then I stumbled into the workunit message logs for the failed W/Us on this website and found a message something like "cannot allocate 1199 mb" in each one. So I thought about it for awhile and decided to suspend all but 3 W/Us, letting only 3 process at any given time. This brought my memory usage down into the 55% range. Workunits began to finish successfully. Not all, but more than failed. I now also limit the number of 16e units to 2 simultanously as they seemed to be the memory hogs. So I will try to finish what I have in my cache. But I feel there is more work needed with memory management before this project is ready for prime time.

I hope this will help someone else out. Dave
ID: 774 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
idahofisherman
Avatar

Send message
Joined: 5 Sep 09
Posts: 1
Credit: 447,182
RAC: 1,157
Message 786 - Posted: 27 Jan 2012, 5:14:23 UTC
Last modified: 27 Jan 2012, 5:17:06 UTC

It seems that all the NFS WUs are getting compute errors of 1 and terminating, but as you can see they are not placing and messages in the message log stating what the error is.

I found the error by looking at the history file in BoincTasks. This seems to be happening on all te WUs. 15 of the last sixteen sent have errored. One is still in process.

History log

NFS@Home 1.09 16e Lattice Sieve S2m1061c_371514_0 00:05:14 (00:03:05) 26-01-2012 09:04 PM 26-01-2012 09:06 PM Reported: Computation error (1,) BARBRAS-LAPTOP

message log

44477 NFS@Home 26-01-2012 08:39 PM Starting task S2m1061c_371514_0 using lasievef version 109
44495 NFS@Home 26-01-2012 08:44 PM Computation for task S2m1061c_371514_0 finished
44497 NFS@Home 26-01-2012 08:44 PM Started upload of S2m1061c_371514_0_0
44498 NFS@Home 26-01-2012 08:44 PM Finished upload of S2m1061c_371514_0_0
44503 NFS@Home 26-01-2012 08:46 PM Sending scheduler request: To fetch work.
44504 NFS@Home 26-01-2012 08:46 PM Reporting 1 completed tasks, requesting new tasks for CPU
ID: 786 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
[XTBA] Raferty

Send message
Joined: 24 Oct 09
Posts: 2
Credit: 1,600,142
RAC: 0
Message 787 - Posted: 31 Jan 2012, 3:46:07 UTC

I experience lots of trouble with your project since a few months ...

first some WUs use lots of memmory 4 of them use 12Gb !!!!
note that it is far from 1Gb each ...

Now I put an AMD 4000+ on your project it is realy stable !
(used on primegrid with intensive WUs without any error)

tested it survey it and conclued that the trouble is not by anyway from this computer !

now look at that :http://escatter11.fullerton.edu/nfs/results.php?hostid=22878&offset=0&show_names=0&state=5&appid=

all the 16 lattice units turn to error in around 4 minutes .
I aborted some of them to keep the machine working on other units .

I thought it would be good to stop working on them but the description of the WU in the option page is not clear enough to make me able to stop those units .

hope you'll find a solution

excuse my english (I'm french)
ID: 787 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Greg
Project administrator

Send message
Joined: 26 Jun 08
Posts: 640
Credit: 433,775,088
RAC: 344,736
Message 790 - Posted: 31 Jan 2012, 21:02:22 UTC - in response to Message 787.  

We are finishing up 2,1061- now, which is the largest number we have ever done, and when completed, a record for the largest SNFS factorization ever. I am not surprised that the sieving is straining resources. I would recommend disabling lasievef sieving for a few weeks until we move on to the next number.
ID: 790 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
[XTBA] Raferty

Send message
Joined: 24 Oct 09
Posts: 2
Credit: 1,600,142
RAC: 0
Message 791 - Posted: 2 Feb 2012, 18:54:09 UTC - in response to Message 790.  

ok thank you for the answer .

I tried some "16e" units on my Q6700 on which I added 4GB and it was able to finish them ...

ID: 791 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Questions/Problems/Bugs : Error in Computing


Home | My Account | Message Boards