Interesting Folding Observation

Discussion in 'Distributed Computing' started by Sedulous, Mar 30, 2003.

  1. macrumors 65816

    Sedulous

    Joined:
    Dec 10, 2002
    #1
    If you go to the folding at home stat page, there is an option to view stats by OS.

    Doing a little math:

    Windows: 337,558 WU / 81,779 processors=
    4.13 WU/processor

    UNIX: 32,129 WU / 4413 processors =
    7.28 WU/processor

    Mac OS X: 23,198 WU / 2551 processors =
    9.09 WU/processor

    Does this tell us something?
     
  2. macrumors 65816

    Joined:
    Aug 20, 2001
    #2
    Pretty much agrees with what I've seen with folding. I've got a Linux dual P3-Xeon 700 w/1MB L2 and it always beats my dual G4-800. Slower clock speeds, slower RAM, less chip cache, and it still wins. I bet one of those dual G4-1.2 cards could equalize things.
     
  3. Moderator emeritus

    edesignuk

    Joined:
    Mar 25, 2002
    Location:
    London, England
  4. macrumors 601

    zimv20

    Joined:
    Jul 18, 2002
    Location:
    chicago
    #4
    ????

    i'm not seeing how those numbers give any indication of folding speed.

    what it tells me is that the mac users tend to stick w/ folding longer (more WU-production bang per processor buck).
     
  5. Moderator emeritus

    Joined:
    Jun 25, 2002
    Location:
    Gone but not forgotten.
    #5
    The only thing I would say is that Mac users are more likely to fold 24/7 than others are.

    The PC users are more likely to play games.

    It used to be that there were more active Mac OS X users, as a percentage, than any other o.s.
     
  6. macrumors 604

    MrMacMan

    Joined:
    Jul 4, 2001
    Location:
    1 Block away from NYC.
    #6
    This is true because no one wants to install Os X which RUNS folding, on a >400MHZ computer

    Our per/WU would be alot lower if they has OS 9 compatibality. ;)
     
  7. macrumors 65816

    Joined:
    Aug 20, 2001
    #7
    zimv20:

    Oops, I read that as "hours per unit". :rolleyes: So basically I read what I expected it to say. Stupid stupid.
     
  8. macrumors 601

    zimv20

    Joined:
    Jul 18, 2002
    Location:
    chicago
    #8
    no worries. it was the middle of the night, i was sleepy and scratching my head. "what am i missing?!?!" lol
     
  9. macrumors 68000

    FelixDerKater

    Joined:
    Apr 12, 2002
    #9
    Macs definitely don't whip PCs in folding. My Compaq K6-III @400MHz is at least two or three times as fast at folding as my 500MHz G4.
     
  10. macrumors 604

    MrMacMan

    Joined:
    Jul 4, 2001
    Location:
    1 Block away from NYC.
    #10
    No one said that at all. :rolleyes:
    Pcs are faster than macs in folding because they can't write the tiner core properly.

    all should be good when we get gromac core tho... they just keep pushing it back :mad:

    Yes that is why we have so many mac members helping out our team.
     
  11. Moderator emeritus

    Joined:
    Jun 25, 2002
    Location:
    Gone but not forgotten.
    #11
    No improvements to the Tinker core would suddenly yield 1-to-1 results against a processor double the clock speed. I think people are going to be disappointed that the Gromacs core will not be as fast as expected, even if it will be much faster than Tinker.

    When the Gromacs core is bug-free, they will let it out to the general public, until that time, they need accurate answers.
     
  12. macrumors 65816

    Joined:
    Aug 20, 2001
    #12
    MrMacman:

    Or is it just because G4's and G3's are vastly outpowered by Athlons and P4's?
     
  13. macrumors 604

    MrMacMan

    Joined:
    Jul 4, 2001
    Location:
    1 Block away from NYC.
    #13
    There should only be a 10-30% G4 ~ P4 difference because the Gromac core is Altevec supported, check there web site they have stats.

    Basically there is a poor written code for the whole mac clients w/ Tinker. Plus with good code the gromac gives a better speed boost to the G4.

    :D
     

Share This Page