1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

8th Annual BOINC Pentathlon: 5th May - 19th May 2017

Discussion in 'Team OcUK Distributed Computing Projects' started by hanluc, Apr 9, 2017.

  1. hanluc

    Wise Guy

    Joined: Oct 22, 2010

    Posts: 2,490

    Location: Ratae Corieltauvorum

    9th is pretty good, not as good as 7th though :p
    PDW will you be moving any resources over to the raccoon lovers this year? they seem to be struggling again.
     
  2. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    I had thought about it but I think they got lucky this year and for some of the projects chosen they have actual Raccoon Lovers working on them.
    I don't think anyone turned up last year and for very little effort I got them at least 1 point, can't find last year's stats as it takes me to this year's stats.
    Also, don't think I can remember the password for the email account/ID I used, but here's a link to their POGS team list still showing me as top contributor by RAC...

    https://pogs.theskynet.org/pogs/team_members.php?teamid=10522&offset=0&sort_by=expavg_credit
     
  3. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    right now all my systems are running POGS. I added cosmology to boinc and all tasks this morning got half way through and all failed. I added LHC and got no work for this at all

    downloaded the boinc + virtual machine from boincs website but not sure how that's suppose to work.
     
  4. Biffa

    Man of Honour

    Joined: Oct 30, 2002

    Posts: 15,043

    Location: Surrey

    I'm going all-in on Einstein, currently at about 75%, should be at 100% tonight
     
  5. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    Were the failed tasks on your big machines or all machines ?
    What were the error messages reported for those tasks on the website in your account ?
    Did you choose just one type of work (which?) or all of them ?

    For LHC you probably just want to try Atlas so de-select the others in your preferences and create an app_config file to limit how many run and how many cores the multi-threaded app uses.

    You just need VirtualBox downloaded and installed, Boinc will use it (if it sees it, I have 1 box that it doesn't work on at the moment because Boinc says it isn't installed when it is !).
     
  6. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    camb_boinc2docker is what is ticked in my cosmo account.

    going to go onto LHC website now and try atlas

    edit- seems cosmology is now downloading on my i5 pc which im using to test it with! using 3 cpus for 1 unit. in the application box it shows it shows (vbox64_mt)
     
    Last edited: May 10, 2017
  7. Biffa

    Man of Honour

    Joined: Oct 30, 2002

    Posts: 15,043

    Location: Surrey

    That means its working
     
  8. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    They have just announced there will be no more Docker work...

    Cosmology@Home: New camb_boinc2docker workunits temporarily disabled
    Hi all, in an effort to keep the server functioning smoothly (its operating pretty close to capacity with all the extra traffic from the Pentathlon) I've disabled creation of new camb_boinc2docker jobs until the contest is over on May 19th. All your in-progress camb_boinc2docker work will finish and be validated, this only means no *new* work for this application will be sent out by the server. If you wish to continue crunching Virtualbox applications (like camb_boinc2docker was), there should be enough planck_param_sims jobs, just make sure you haven't explicitly disabled this application in your preferences. Good luck and happy crunching!

    So you need to have Legacy and/or Planck options ticked in your preferences !
    I found out because I'd run out and sitting idle :mad:
     
  9. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    can somebody tell me what an app cfg would look like for a 32 core system running plank_param one with 28gb ram and the other with 16gb

    it seems my i5 has completed some units on this so ill run with it for now
     
  10. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    Something like this...

    <app_config>
    <app>
    <name>camb</name>
    <max_concurrent>1</max_concurrent>
    </app>
    <app>
    <name>lsplitsims</name>
    <max_concurrent>1</max_concurrent>
    </app>
    <app_version>
    <app_name>lsplitsims</app_name>
    <plan_class>vbox64_mt</plan_class>
    <avg_ncpus>4</avg_ncpus>
    </app_version>
    <project_max_concurrent>1</project_max_concurrent>
    </app_config>

    I removed the Docker stuff as you can't get that now.
    'camb' is the legacy non-VB work.

    'lsplitsims' is the Planck work.
    Each multi-threaded Planck job uses 4 threads (avg_ncpus)

    max_concurrent says only 1 job for each type, project_max_concurrent says only allow 1 job from the project to run at a time.

    You need to tune these settings (how many you run, and how many threads) to your hardware.

    Even running 8 Plancks (on 4 threads each) may be too much for the amount of memory you have.
    You need to monitor memory usage to see how many you can do at a time, either pad out remaining cores with other low memory work or leave them idle.
     
  11. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    ok ill see how it goes with the cpu set at 4 then if its to much should I knock it down to 2?
    if I change the app cfg will it adjust the WU's that I already have or will I need to remove them and download some more
     
  12. MGP

    Mobster

    Joined: Oct 24, 2004

    Posts: 2,510

    Location: Surrey

    Sprint project is Yoyo@home

    13 May 00:00 UTC to 16 May 00:00 UTC
     
  13. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    No, you will be able to run 4 threads at a time per job but you may not be able to run 8 jobs at the same time on the big machine (28Gb).

    Try 4 jobs (using 4 threads each) and see what memory is like, if memory is ok try running 5 or 6 jobs (4 threads each).
    You may not get 4 jobs (4 threads each) on the 16Gb machine, not sure. I run out of cores/threads before I run out of memory !
     
  14. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    ok got 4 jobs with 4 threads each going and heres a screen shot. says its only using 6.2gb ram?
    [​IMG]
     
  15. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    Was that memory usage when they have been going for a few minutes ?
     
  16. ozaudio

    Mobster

    Joined: Jan 23, 2010

    Posts: 2,923

    Location: St helens

    7.2 mem usage

    my account says its completed 2 work units so far.
     
  17. Xez

    Wise Guy

    Joined: Jun 24, 2005

    Posts: 2,024

    Location: Lincolnshire

    Sorry guys, been admitted to hospital earlier hours of this morning so ive had to turn off my main computer that has the GPU. My microserver is using its CPU though to continue just slowly...
     
  18. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,244

    No worries, some things are more important than a fun yearly challenge, hope everything goes okay and you get back home soon.
     
  19. Biffa

    Man of Honour

    Joined: Oct 30, 2002

    Posts: 15,043

    Location: Surrey

    Hope it's not too serious Xes
     
  20. MGP

    Mobster

    Joined: Oct 24, 2004

    Posts: 2,510

    Location: Surrey

    Anyone got any extra power for Cosmology, else we are going to lose a slot in the next day or so.
     


Share This Page