We Made Red!!!!!

Discussion in 'Distributed Computing' started by MacBandit, Feb 13, 2004.

  1. MacBandit macrumors 604


    Aug 9, 2002
    Springfield, OR (Home of the Simpsons)
    I just checked the folding.extremeoverclocking.com site and it shows us at 32,000+ points in the last 24hours. I wish I could say I contributed a large part of that but lately I haven't had as many machines to do my dirty work.

    Good job to everyone though.
  2. Mr. Anderson Moderator emeritus

    Mr. Anderson

    Nov 1, 2001
    Its short lived, though. If you check the team stats we had a 16k day the day before. Looks like there was a server fart and all the WUs that didn't get uploaded previously, swarmed in and gave us a good bump up.

    Now if we can do it 2 days in a row I'll have to come up with another explanation :D

  3. bousozoku Moderator emeritus

    Jun 25, 2002
    Gone but not forgotten.
    I've long-running, high point value WUs running now. I got rid of a 70.77 point WU yesterday. Perhaps, it's similar all over and we'll have more peaks and valleys for a while, averaging us back in the orange.
  4. MacBandit thread starter macrumors 604


    Aug 9, 2002
    Springfield, OR (Home of the Simpsons)
    Agreed I have been getting the large work units as well. I like them personally I can turn out more points with the large ones in a day then I can with multiple short work units.
  5. shemp9999 macrumors regular


    Dec 29, 2003
    that was 1.5 days worth of wu.

    i'm doing over 1200 a day now, and it looks like that might stick - thanks to that last dual G5. :D

    i've been surrounded by 48 GHz of xeon power (12x2x2GHz) all week. been fighting all kinds of urges to add them to my queue. ten minutes is all i'd need :p

    but i just don't feel right using machines outside my general department. i could get away with using hundreds of GHz - for a while. that reminds me of the ethics question here the other day...my answer would be - i'd sell a machine as is, daemons be damned. wipe it if you want, luser. but i would.

    back on message, these are pipeline training boxes. all i need is some inquisitive new kid asking what the proc heavy FAHCore processes are...i fear they wouldn't grok the idleness of it all. so i shan't get greedy.

    would't it be cool to double the points, though.

    hey; maybe i'll figger out how to launch at 6pm and kill at 9 am? my ethics might fit that under their frock. :D

    would these be the appropriate crontabs?

    15 18 * * * cd ~/.fah/fah1; ./fah4 -local -advmethods -forceasm
    15 18 * * * cd ~/.fah/fah2; ./fah4 -local -advmethods -forceasm
    45 8 * * * killall fah4
  6. MacBandit thread starter macrumors 604


    Aug 9, 2002
    Springfield, OR (Home of the Simpsons)
    No as stated by the site that was 24 hours worth of production. Just so happens we had a couple of large hours at 2 or 3 in the morning that drove our 24 hour total way up. If you took the normal day though from midnight to midnight it was lower but from 3-3 it was 30k+.
  7. jbeetz macrumors member

    Jan 8, 2004
    That's close, but with cron I've found that you have to be VERY explicit about the path. Replace the ~ with /Users/username. And be sure to redirect the output to /dev/null and background the process. I don't know where cron puts stdout if you don't send it to /dev/null. Also, to let machines run non-stop through weekends, replace the third * with 1-5 on each line.

    so try this...

    15 18 * * 1-5 cd /Users/username/.fah/fah1; ./fah4 -local -forceasm > /dev/null &
    15 18 * * 1-5 cd /Users/username/.fah/fah2; ./fah4 -local -forceasm > /dev/null &
    45 8 * * 1-5 killall fah4

    I tested it by setting the launch time a minute ahead of the time of day then making sure the processes launched. Took some trial and error but it does work!

Share This Page