Announcement

Collapse
No announcement yet.

Genome@home has stopped distributed wu and ends 15. April 2004.

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Genome@home has stopped distributed wu and ends 15. April 2004.

    In case someone still crunched genome@home around here, these is the latest news:

    March 8, 2004
    It is with the deepest gratitude that we wish to thank all those who have contributed over the years to the Genome@home project. It has accumulated a large database of protein sequences, which will be used for important scientific purposes for many years by our group and many others throughout the world. Like all good things, however, its time has come. Over the last year we have kept Genome@home running on a very low budget by sharing much of its client infrastructure with Folding@home, but several months from now its funding will no longer be available at all.

    For this reason, the Genome@home project will be terminated. While the servers will stop distributing new work units on March 10, they will remain up until April 15 to receive the final results.

    Our sister project Folding@home remains on very solid financial ground. While it requires a more frequent connection to the Internet and it has a somewhat different scoring system for those who enjoy the competition, we would highly recommend it. Your continued donation of computer time to Stanford University in the Folding@home project will always be, as it has been with Genome@home, truly appreciated.

    With great thanks,

    Vijay Pande, Stefan Larson, & the rest of the Pande Grou

  • #2
    You could always move over to http://www.climateprediction.net/
    DM says: Crunch with Matrox Users@ClimatePrediction.net

    Comment


    • #3
      Or join the fold.

      Comment


      • #4
        Yeah, join distributedfolding.org, Team Matroxusers.
        RC Agent
        AMD Athlon 64 X2 5000+ Brisbane 2.6GHz, MSI 785GT-E63, 6 GB(2x1GB, 2x2GG) DDR2 800 Corsair XMS2, Asus EAH4850 TOP
        AMD Athlon 64 X2 7750 Kuma 2.7GHz, ASRock A790GXH/128M BIOS 1.7, 4 GB(2x2GB) DDR2 800 Corsair XMS2, Gigabyte HD 6850 1GB DDR5
        AMD Phenom II X6 1045T 2.7GHz, Asus M5A99FX Pro R2.0 BIOS 2501 , 8GB(2x4GB) DDR3 1866 CL9 Crucial BallisticX(BLT4G3D1869DT1TX0) , Sapphire HD7870 2GB GDDR5 OC, Seasonic 850w powers supply

        Comment


        • #5
          Well, MURC has teams in UD, folding@home, DF, CPDN & Seti@home.
          Genome@home had 5 active crunchers the last week.


          The easiest & most stable to run is seti@home, and is the largest MURC-team both in members and active.

          Folding@home has the best client appart for seti@home. There's time-limits on wu, but saw something about it's now possible to get wu without time-limit...

          UD has terrible stats-system & problematic client, and AFAIK also has time-limit. (too long since last checked.

          DF client needs to use upto 128 MB memory to run effective, credits different wu the same amount even crunch-time is wastly different, can continue to crunch as long as there's available space on hd but if for some reason misses the currently unknown switch-over to a new client/wu by 2 days will not get any credit.

          CPDN is by far the most demanding in hd-space & crunch-times, there even a new computer can use a month to crunch a wu. The client is 7 MB in download, but wu is very small (1 KB or something). The result is around 5 MB, but with the long crunch-times this project is probably the least demanding on your internet-connection.
          Every 10800 time-step a "trickle" is generated, just a small info-package on progress & status that is uploaded next time you've connected, and your stats is based on this so should be updated every 1-2 day depending on speed. CPDN is not recommended for users that only have machine on 1-2 hour/day.
          CPDN has many connection-problems there often must do a manual upload, and the client or wu often crashes leading to a short-run. Based on this, my suggestion to someone that not already is running CPDN is to wait till the BOINC-version is released later this year.

          For all with HT or multi-cpu-systems, neither UD nor CPDN can run more than 1 instance so you'll need to run another DC-project alongside to utilize all cpu-power...

          Of course, it's also possible to help beta-test BOINC.

          Comment


          • #6
            Originally posted by Rattledagger
            [B]. . .
            DF client needs to use upto 128 MB memory to run effective, credits different wu the same amount even crunch-time is wastly different, can continue to crunch as long as there's available space on hd but if for some reason misses the currently unknown switch-over to a new client/wu by 2 days will not get any credit.
            . . .
            This was an enhancement they came up with later for speed. It is easilly changable.

            Comment


            • #7
              I'd have to agree with Rattledagger actually - if you aren't crunching with Matroxusers@CPDN yet, then it's probably best to wait for the BOINC client...
              DM says: Crunch with Matrox Users@ClimatePrediction.net

              Comment


              • #8
                Originally posted by High_Jumbllama
                This was an enhancement they came up with later for speed. It is easilly changable.
                I've actually not benchmarked DF with/without this switch, but from the faq it looks like nearly doubling of the production then enabled... So to run DF effectively you'll need to use upto 128 MB memory.

                That being said, I've for different amount of time crunched all of these project, but is currently only running seti@home & BOINC.

                Comment


                • #9
                  I had considered the client project but the fact that info is lost if you do not finish a run makes me think the programming decision makers do not know what they are doing. Correct me if I am wrong on this. I remember this was the case in the beginning.

                  BOINC will be good. There's no disagreement about that.
                  Last edited by High_Jumbllama; 13 March 2004, 16:49.

                  Comment

                  Working...
                  X