| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Need help understanding errors and Boinc behaviour |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 11
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello,
Recently I visited by chance the page "Results Status" within my profile and I would really appreciate if someone help me making some sense out of it. My Results Status includes 29 pages of information. Here are my questions: - From page 17 to 29 (about 180 different results listed, 15 per page) basically all my results are reported as "Error". What does that mean? - From page 1 to 17 I get about 255 different results listed as "in progress". I'd say that it's extremely unlikely that my machine will process all those tasks in time. Why do I have so many tasks "in progress"? So I decided to check BOINC. What I found in the messages tab was a frenzied activity for today, hundreds of tasks being aborted and downloaded. I don't understand what's happening. I must say that I *never* bother to keep an eye on BOINC, I just leave it alone, so I really don't know if this has been the standard behavior around here for the last couple years or if it's a one time thing. Things out of ordinary here: - My CPU was undervolted for a reduction on core temperature. I did this years ago, and tried to be sure that it would lead to no computation errors by running prime. The computer has been working perfectly ever since. - Recently I increased the hard disk space available for BOINC. I did this after reading in the Messages tab that some projects demanded more space. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm trying to paste BOINC's message history here. I can preview the post fine, but I try to publish it I get this message from the server:
Error executing SQL in PostDAOImplJDBC.create. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
An ineligant way of telling that the post has like more than 32K or 36K characters of text. If there's a lot of repetition in the log you wish to share, just post top 40 / bottom 20 lines.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Rob, do you mind cheking the log here?
www.fabiobustamante.com.br/downloads/log-boinc.txt I wouldn't know how to filter whats important in it... |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That looks dismal.
1) A power machine 2) System date correct (important 3) The OpenCL part (a second line of the ATI card is missing), suspecting need to install the ATI-SDK is required. Personally, if the client cant confirm to WCG that OpenCL is loaded, no HCC-GPU tasks should be assigned, but what is serially aborted *too* are CPU tasks for sciences such as DSFL, to the point that the quota is getting exhausted or the feeders are getting constipated with all those aborted jobs that have to be reissued to reliable clients. The last series of 1 task assigned at the time suggests a disk full, but it is not per the top of the log. Was having discussion with the techs in the situation room on another matter and feel this is somehow related, so here goes: Calling Techs This from a quick scan |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ok, Rob, I'll wait for the tech's comments.
Another out of ordinary thing that happend here: ever since I updated my GPU a couple years ago I never tried to crunch with it. So a few days ago I decided to see how I worked and activated GPU usage. Apparently BOINC managed to complete one task using my GPU, but fan noise got so high I decided to disable it again. I don't know how it was set initially, but I noticed I left "Use GPU while computer is in use" ticked in preferences and blocked GPU by checking "Use GPU never" in the Activity menu. Now I unchecked GPU processing in preferences too. Regarding the 12 pages of results listed as "error" in my Results Status, is this all part of the same problem? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
A way to ensure that GPU tasks never start if they erroneously arrive, even if deselected in the device profile, is to set a ludicrous wait time for idle GPU computing to resume. I've entered 20000 minutes in my prefs. Or in general to tell all projects there is no GPU, to insert the <no_gpus>1</no_gpus> in the cc_config.xml, <options> section,
But, no OpenCL is installed per your log, which is an equal sign to me not to assign GPU tasks and let the server tell you there is none assigned due this reason. |
||
|
|
OldChap
Veteran Cruncher UK Joined: Jun 5, 2009 Post Count: 978 Status: Offline Project Badges: 
|
fabiopb:
----------------------------------------From past experience I wonder if you have your cache set too large? This can present as error when work is too late or not started in time In Boincmanager go to tools > Computing preferences > Network usage and set minimum work buffer for maybe 1 day and additional work buffer for half a day. Are you wanting to run GPU wu's? If so you need to say so in your profile for that rig. EDIT: just read your later post  Have you restricted your Machine to run just 4 cores or perhaps you have hyperthreading turned off?? Like Rob this is what I wonder after having a quick look at your log.  [Edit 1 times, last edit by OldChap at Mar 18, 2013 6:19:20 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My work buffer was in 2 days. I set it to half a day like you suggested.
Yes, HT is off in my machine. This lead to slightly cooler temps, also I could use all my processing power with 4 WCG tasks instead of 8, saving half the RAM. No, I don't intend to run on GPU because as far as I could see there's no GPU throttle, and my video card apparently starts making a terrible noise when used 100%. I think it's a good idea to update BOINC here. I'm running version 6.10.58 and the latest version is 7.0.28... |
||
|
|
OldChap
Veteran Cruncher UK Joined: Jun 5, 2009 Post Count: 978 Status: Offline Project Badges:
|
With GPU it might be possible to find the fan speed that is acceptable and fix it at that speed using Afterburner or Trixx then observe the temperature when it runs a work unit. If you find that temperature acceptable too then all is well, if not then it may be possible to reduce the core speed of the card some and thus the voltage to get into a good temperature zone.
----------------------------------------My experience with Boinc 7.0.55 has been a good one. I hear that the current 7.0.56 beta may become the next recommended. There may be other advantages too that are incorporated in this one [Edit 1 times, last edit by OldChap at Mar 18, 2013 7:26:39 PM] |
||
|
|
|