Can Anyone Answer Questions/Problems Here?


Advanced search

Message boards : SZTAKI Desktop Grid : Can Anyone Answer Questions/Problems Here?

AuthorMessage
KAMCOBILL
Send message
Joined: Feb 26 06
Posts: 26
Credit: 134,972
RAC: 2
Message 3873 - Posted 25 Aug 2006 0:27:30 UTC

    Message 3527 - Posted 14 Jul 2006 14:57:18 UTC [Edit this post]

    I think there is a real problem with work units. WUs have 3 day deadline and are running 100+ hours. Doesn\'t seem possible to complete them without getting No reply and credit. Here\'s a few post on subject: http://szdg.lpds.sztaki.hu/szdg/forum_thread.php?id=396

    WU 138 and 266 are the two I have. I suspending them after deadline at 20% or less. I have one another with a 7 day which is better, but it\'s been running 12 hours with 0% progress and 101:13:02 hours still left to go.

    I\'m running other projects of 10 minutes duration with a 7 day deadline. Could someone look into this problem? It looks like it going to cause some detaching.

    I\'m going to detach all but 1 node just to see if it can meet the deadline.

    Thanks

    Bill
    ____________


    ID: 3527


    Message 3841 - Posted 19 Aug 2006 18:22:19 UTC [Edit this post]

    They finally crashed. But Whoa! I finally go a result through :-)

    Not sure what the errors means: Too many total results

    Maybe someone can help explain? http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=24936

    Hopefully it\'s not too many claimed credits. If it is look at this one: http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=40436

    I re-attached all nodes back to the project and am getting really long workunits. WU 40436 is the smallest that I have running now. I running 2.6 Ghz PCs full out and the time to finish is still climbing at about 15% per day.

    One started out at ~105 hours to finish, has been running for ~90 hours and now has ~114 hours to finish.

    Another one started at started out at ~76 hours to finish, been running for ~80 hours and has 117 hours to go.

    All PCs running Project has stopped fetching and using earliest deadline becaused they are overcommited.

    Hope they make it through. Two out of 4 has so far. Fingers are crossed.

    Happy crunching

    Bill


    ____________


    One started out at ~105 hours to finish, has been running for ~90 hours and now has ~114 hours to finish. Now it\'s been running for about ~105 and has ~114 to finish. Deadline is August 24.

    Another one started at started out at ~76 hours to finish, been running for ~80 hours and has 117 hours to go. Now it\'s been running for ~110 and has ~138 to finish. Deadline is Aug 26.

    Does anyone know if the Server is giving a week grace still? If no, I don\'t think deadline can be met.

    Thanks

    Bill
    __________________________________

    One started out at ~105 hours to finish, has been running for ~90 hours and now has ~114 hours to finish. Last it\'s been running for about ~105 and has ~114 to finish. Now it\'s benn running ~127 and ~110 to finish (40 %). Deadline is August 24.

    Another one started at started out at ~76 hours to finish, been running for ~80 hours and has 117 hours to go. Last it\'s been running for ~110 and has ~138 to finish. Now it\'s benn running ~141 and ~160 to finish (20.6 %). Deadline is Aug 26.

    ____________


    Now it\'s benn running ~148 and ~106 to finish (46.5 %). Deadline is August 24.

    Now it\'s benn running ~165 and ~177 to finish (21.3 %). Deadline is August 26.

    With this progress, Deadline could be met in November hopefully.



    ____________


    ~193 running 195 hours to finish 22.3% completed. Deadline Aug 26th



    Deadline over 103 hours remain to finish 50.1% completed
    http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=39327

    Will I still get credit if I finish them?
    ____________


    Can Someone give me an answer?

    ____________


    ID: 3872

    ____________

    Stargazer257
    Send message
    Joined: Feb 8 06
    Posts: 5
    Credit: 339,461
    RAC: 0
    Message 3900 - Posted 28 Aug 2006 4:32:06 UTC

      I\'ve the same question as KAMCOBILL has.

      I\'ve got 2+ days into WU 39584 http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=39584 and at the current rate it \"might\" just beat the reporting deadline of 19 Sept. That is assuming I don\'t run anything else AND don\'t turn off the machine.

      So, should I ditch it? Or should I continue on?

      Will the rate of completion for this WU increase or should I expect it to continue at about the same pace?

      If I put 20+ days into this work for naught, I dare say that I would be unlikely to work SZTAKI again.
      ____________


      Join Us! - Click the Sig!

      Dronak
      Avatar
      Send message
      Joined: May 25 06
      Posts: 2
      Credit: 32,450
      RAC: 0
      Message 3903 - Posted 28 Aug 2006 16:31:17 UTC - in response to Message 3900.

        I\'ve got 2+ days into WU 39584 http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=39584 and at the current rate it \"might\" just beat the reporting deadline of 19 Sept.

        So, should I ditch it? Or should I continue on?


        It looks like that work unit already has three client error results. This happened with one of my units, too, and I aborted it. It didn\'t look like spending tons of time on a unit that would very likely fail (3 errors already, and 4 is the max) was worth it. I then blocked new work requests so that I wouldn\'t have to deal with this again. I\'m waiting for information about fixes to the recent problems like this one. So if I were you, I\'d abort it. That\'s just my opinion though.
        ____________

        NJMHoffmann
        Send message
        Joined: May 19 06
        Posts: 18
        Credit: 10,221
        RAC: 0
        Message 3905 - Posted 28 Aug 2006 19:03:36 UTC - in response to Message 3903.

          Last modified: 28 Aug 2006 19:03:51 UTC

          It didn\'t look like spending tons of time on a unit that would very likely fail (3 errors already, and 4 is the max) was worth it.

          And with total results set to 5, the chance to get 3 good results is about zero. I abort everything, that already has 3 errors, if 3 good results are needed. (And about 90% of the WUs I got recently looked like that.)

          ____________

          Stargazer257
          Send message
          Joined: Feb 8 06
          Posts: 5
          Credit: 339,461
          RAC: 0
          Message 3907 - Posted 29 Aug 2006 5:01:15 UTC

            Thanks for the feedback. WU ditched. Will give SZTAKI another chance, but will check each WU for similar scenarios...

            Thanks again!
            ____________


            Join Us! - Click the Sig!

            VAIO GRX616
            Send message
            Joined: Jan 9 06
            Posts: 1
            Credit: 11,927
            RAC: 0
            Message 3911 - Posted 30 Aug 2006 13:23:51 UTC

              I leave my machines doing their thing and don\'t interfere. I have a WU (ab094285-a60b-418b-a548-bec0c0d6271d_6732e8f3-cbc2-4238-b3cf-e58246b72ce1_1931) that has \'Too many results\' and it seems I get no credit for it - after 100+ hours on a fast machine! This needs sorting ASAP!!!

              Very annoyed >:-(
              ____________

              AlphaLaser
              Send message
              Joined: Jul 26 06
              Posts: 2
              Credit: 1,880
              RAC: 0
              Message 3913 - Posted 1 Sep 2006 19:42:39 UTC

                Last modified: 1 Sep 2006 19:52:11 UTC

                I also have a problem with the high error rates and low # of max results. See this WU. This is the first wu I\'ve managed to finish. However two hosts have returned errors and if the next one also errors out it seems that 105 hours of work will be lost. :(

                Just browsing randomly through other WU\'s with slightly different ID (example #39121, #39122, #39123, #39124) it seems that a very, very large number of WU\'s are suffering this problem. Not fair to the people who are managing to return results on time...
                ____________

                Profile Ananas
                Send message
                Joined: Jul 12 05
                Posts: 222
                Credit: 665,833
                RAC: 0
                Message 3921 - Posted 4 Sep 2006 0:05:47 UTC

                  The SZTAKI project is currently a pile of crap. Two valid WUs per day, a real waste of money to put electricity in here.

                  Disable downloads and wait until Adam understood that he needs to fix stuff before he can expect people to crunch for SZTAKI.

                  Other projects will be happy for the CPU cycles and when SZTAKI has a new and tested(!!!) project client, re-enable downloads here.


                  I really do not understand why Adam does not close the project while it is in this state. He could work on it without having to care for credits and timeouts and so.

                  Having it open to the public under these circumstances does way more damage than closing it for the time he needs to get things sorted.

                  Profile Rebirther
                  Avatar
                  Send message
                  Joined: Jul 12 05
                  Posts: 81
                  Credit: 15,472
                  RAC: 0
                  Message 3922 - Posted 4 Sep 2006 7:33:28 UTC

                    You are right Ananas. Before all has began here with the new app, it was totally the wrong way. Why you need 7 WU generators while WUs running days? Split all units in some smaller pieces for faster computing, \"RCN\" did it so and all users are happy to crunch a WU in max. 4h instead max 900h!

                    But if the situation is staying so then the project runs into dead...!

                    Profile [b@h] tomcat
                    Send message
                    Joined: Feb 4 06
                    Posts: 20
                    Credit: 21,462
                    RAC: 0
                    Message 3924 - Posted 4 Sep 2006 14:22:52 UTC - in response to Message 3922.

                      Last modified: 4 Sep 2006 14:26:37 UTC

                      I agree with Ananas too.
                      for further development it may be useful to make a \"Test-Project\" for testing new applications, like ThEfT sugested (General suggestion ...)
                      ____________
                      tomcat
                      - member of BOINC@Heidelberg

                      Profile Nightbird
                      Forum moderator
                      Avatar
                      Send message
                      Joined: Jul 12 05
                      Posts: 920
                      Credit: 114,924
                      RAC: 0
                      Message 3925 - Posted 4 Sep 2006 18:40:44 UTC - in response to Message 3924.

                        I agree with Ananas too.
                        for further development it may be useful to make a \"Test-Project\" for testing new applications, like ThEfT sugested (General suggestion ...)

                        Your idea is not new : Honza has (already) suggested the same idea, 2 months ago.

                        ____________

                        Profile Nightbird
                        Forum moderator
                        Avatar
                        Send message
                        Joined: Jul 12 05
                        Posts: 920
                        Credit: 114,924
                        RAC: 0
                        Message 3927 - Posted 4 Sep 2006 19:04:11 UTC - in response to Message 3921.

                          Last modified: 4 Sep 2006 19:07:47 UTC

                          The SZTAKI project is currently a pile of crap. Two valid WUs per day, a real waste of money to put electricity in here.

                          Disable downloads and wait until Adam understood that he needs to fix stuff before he can expect people to crunch for SZTAKI.

                          Other projects will be happy for the CPU cycles and when SZTAKI has a new and tested(!!!) project client, re-enable downloads here.

                          I really do not understand why Adam does not close the project while it is in this state. He could work on it without having to care for credits and timeouts and so.

                          Having it open to the public under these circumstances does way more damage than closing it for the time he needs to get things sorted.

                          I can\'t speak for Adam but i say that i suspended Sztaki on 2/3 machines.
                          I\'m waiting for the end of this wu on the third machine.

                          ____________

                          Post to thread

                          Message boards : SZTAKI Desktop Grid : Can Anyone Answer Questions/Problems Here?


                          Home | My Account | Message Boards


                          Copyright © 2017 SZTAKI Desktop Grid