1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

TPU's WCG/BOINC Team

Discussion in 'World Community Grid (WCG)' started by Chicken Patty, Feb 20, 2009.

  1. [Ion]

    [Ion] WCG Team Assistant

    Joined:
    Sep 15, 2009
    Messages:
    12,445 (6.48/day)
    Thanks Received:
    12,364
    Location:
    Raleigh, North Carolina, United States
    Awww yes! :D :D :D :D
    I've picked up some of them on about half a dozen systems now. This means I should get a Gold badge in Beta this weekend! :D :D
     
    Crunching for Team TPU
  2. TRWOV

    TRWOV

    Joined:
    Aug 11, 2011
    Messages:
    3,739 (3.05/day)
    Thanks Received:
    2,329
    Location:
    Mexico
    If things go well this Christmas I might have 12 more cores crunching by next year's February. :D Daddy wants daily pie.

    Arms race, GO!
     
    Chevalr1c, Vinska and Arjai say thanks.
    Crunching for Team TPU
  3. TRWOV

    TRWOV

    Joined:
    Aug 11, 2011
    Messages:
    3,739 (3.05/day)
    Thanks Received:
    2,329
    Location:
    Mexico
    Adding 8 more cores. Waiting for the definitive memory and cooler to arrive but in the mean time this will have to do:

    [​IMG]
     
    Crunching for Team TPU
  4. [Ion]

    [Ion] WCG Team Assistant

    Joined:
    Sep 15, 2009
    Messages:
    12,445 (6.48/day)
    Thanks Received:
    12,364
    Location:
    Raleigh, North Carolina, United States
    More cores is always better :D
    What model FX8 is that? And why two GPUs?
     
    Crunching for Team TPU
  5. james888

    james888

    Joined:
    Jun 27, 2011
    Messages:
    4,737 (3.72/day)
    Thanks Received:
    1,944
    Carefull of those corsair CX series. I had two blow up on me. The first one and its replacement from RMA.
     
    Crunching for Team TPU
  6. TRWOV

    TRWOV

    Joined:
    Aug 11, 2011
    Messages:
    3,739 (3.05/day)
    Thanks Received:
    2,329
    Location:
    Mexico
    8350. 6450 for display, 7770 for Folding.



    It's a V2, don't know if that makes a difference.
     
    Crunching for Team TPU
  7. Norton

    Norton WCG-TPU Team Captain

    Joined:
    Dec 21, 2011
    Messages:
    9,110 (8.32/day)
    Thanks Received:
    21,422
    I had issues with the CX430 getting overloaded with a 965BE but have been running a CX500 for over a year 24/7 with an FX-8150 or an overclocked FX-6200.

    I'm guessing it's the luck of the draw with them :rolleyes:

    Both of them were/are V2's btw
     
    Crunching for Team TPU More than 25k PPD
  8. james888

    james888

    Joined:
    Jun 27, 2011
    Messages:
    4,737 (3.72/day)
    Thanks Received:
    1,944
    Mine was the 600w. I could not find any kind of version number on the box or the psu itself.
     
    Crunching for Team TPU
  9. (FIH) The Don

    (FIH) The Don

    Joined:
    Aug 11, 2009
    Messages:
    6,823 (3.49/day)
    Thanks Received:
    2,910
    Location:
    Viking Land
    im back for a while, with the 3570k and my gpu if it can use it :)
     
  10. mstenholm

    Joined:
    Nov 7, 2009
    Messages:
    2,136 (1.14/day)
    Thanks Received:
    1,272
    Welcome back. Sorry no GPU work. If there was we held the top spot ;)
     
    Crunching for Team TPU More than 25k PPD
  11. (FIH) The Don

    (FIH) The Don

    Joined:
    Aug 11, 2009
    Messages:
    6,823 (3.49/day)
    Thanks Received:
    2,910
    Location:
    Viking Land
    bleh, why havent they figured it out yet, its been like a year since i last crunched, should be fixed by now i would say
     
  12. mstenholm

    Joined:
    Nov 7, 2009
    Messages:
    2,136 (1.14/day)
    Thanks Received:
    1,272
    Yeah it makes you wonder if they care about other than getting their WUs processed. There is a big difference in energy efficiency between 80 WU/hour at 400 W and 6/hour at 140 W. I give them until the end of this year to do the right thing.....
     
    Crunching for Team TPU More than 25k PPD
  13. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    Yes! I also give them to the end of the year! :mad:
    Because if they don't...


    ...I will do nuttin', as there's nothing I can do about it :ohwell:
     
    Crunching for Team TPU
  14. [Ion]

    [Ion] WCG Team Assistant

    Joined:
    Sep 15, 2009
    Messages:
    12,445 (6.48/day)
    Thanks Received:
    12,364
    Location:
    Raleigh, North Carolina, United States
    You do have to realize, that GPU architecture is very different from x86_64 and it's quite difficult to adapt their work. They've released things saying that they've tried experimenting with other projects and GPUs, and some saw no benefit vs solely CPU & other actuallys aw worse performance. I'm really hoping that they can get MCM on GPUs--since it's by the same research team it's somewhat more likely, I think. As much as I would also love more GPU work, we'll just have to wait for now.
     
    Chevalr1c says thanks.
    Crunching for Team TPU
  15. james888

    james888

    Joined:
    Jun 27, 2011
    Messages:
    4,737 (3.72/day)
    Thanks Received:
    1,944
    GPU architecture is very good at doing very simple tasks very very extremely quick such as graphics. They can do more complex stuff too though as well all know, but there strength is in simple and quick.
     
    Crunching for Team TPU
  16. xvi

    xvi

    Joined:
    Nov 10, 2006
    Messages:
    2,231 (0.75/day)
    Thanks Received:
    1,558
    Location:
    Washington, US
    ..and parallel.

    (..which is kind of the definition of distributed computing, but still)

    You could always program your own GPU simulations for them. :laugh:
     
    Last edited: Nov 18, 2013
    Crunching for Team TPU
  17. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    except I have no idea what those simulations actually do and how they work :ohwell:
     
    Crunching for Team TPU
  18. TRWOV

    TRWOV

    Joined:
    Aug 11, 2011
    Messages:
    3,739 (3.05/day)
    Thanks Received:
    2,329
    Location:
    Mexico
    I wonder if they would eventually support the Intel Xeon Phi. If adapting their routines to OpenCL is difficult or impossible I guess they could leverage those Phis more easily.

    Of course, it isn't as if you could buy a Phi on newegg or something :laugh:
     
    Chevalr1c says thanks.
    Crunching for Team TPU
  19. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    Well, if a project has GPU work on Radeons, it is already OpenCL. Since CUDA is nvidia-only. So I assume yes - it would run on Intel Xeon Phi, assuming it is OpenCL work AND that WCG's scheduler isn't too dumb to understand a simple "A Xeon Phi is fine, too".
     
    Crunching for Team TPU
  20. Chevalr1c

    Chevalr1c

    Joined:
    Sep 3, 2010
    Messages:
    3,307 (2.11/day)
    Thanks Received:
    1,484
    Location:
    NL, Europe
    For those who missed it: Darwin (Intel-desk (main rig) reborn) started crunching this afternoon GMT+1. Estimated crunching time needed for the first 4 WUs is 12-16 hours (they are CEP2 units).
     
    Arjai, (FIH) The Don and Norton say thanks.
    Crunching for Team TPU
  21. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    is it just me are is WCG a bit short on FAAH work? I see a lot of my boxen pulling loads of MCM + CEP2 instead of FAAH. Some boxen don't even have FAAH WUs at all...
    (I only got FAAH & SN2S selected + "If there is no work available for the project(s) I have selected above, please send me work from another project.", so it seems FAAH is running short)
     
    Crunching for Team TPU
  22. xvi

    xvi

    Joined:
    Nov 10, 2006
    Messages:
    2,231 (0.75/day)
    Thanks Received:
    1,558
    Location:
    Washington, US
    Just did a manual update and got 20-30 new FAAH WUs. It wouldn't appear that there's any issue for me.
     
    Crunching for Team TPU
  23. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    Aye, I still get them, a lot of them. Don't misunderstand - I did not say I am not getting them anymore.
    But I do seem to get quite a lot of other projects, too. So I believe levels of stored FAAH WUs in the WCG scheduler are somewhat lower than normal so it must be mixing in some other projects in sometimes to maintain steady stored WU levels and such. I dunno... *shrug*
     
    Crunching for Team TPU
  24. [Ion]

    [Ion] WCG Team Assistant

    Joined:
    Sep 15, 2009
    Messages:
    12,445 (6.48/day)
    Thanks Received:
    12,364
    Location:
    Raleigh, North Carolina, United States
    I think you did something wrong; looked at a few of my systems and they're all just getting FAAH WUs.

    :p
     
    Crunching for Team TPU
  25. Vinska

    Vinska

    Joined:
    Jul 23, 2011
    Messages:
    1,420 (1.14/day)
    Thanks Received:
    1,262
    Location:
    Kaunas, Lithuania
    Seems to be at something like "bursts":
    >all get FAAH
    >a burst of other projects for all of my machines asking for WUs at that time
    >all get FAAH again
    rinse and repeat

    Of those new boxen, those that got set up 2nd-to-6th, all got a initial load of all-not-FAAH WUs. So I suppose those "hit" one of those "bursts".

    Since WCG probably has several scheduler servers & thus probably separate scheduler servers for 'Murica and Yurip, it could be that this is for Europe only. Or for my region only, even.
     
    Crunching for Team TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page