Can't get any new wu

Forums › Problems and Bug Reports

candido
candido
Joined: 8 Mar 11
Posts: 11
Credit: 1005637
RAC: 0
18 Oct 2011 19:17:15 UTC
Topic 196024
(moderation:
    )

    since a few days ago I cant get any new jobs
    and I keep geting this message on the event log:

    18/10/2011 20:13:24 | Einstein@Home | see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/4142/4142843

    If we look into http://einstein.phys.uwm.edu//host_sched_logs/4142/4142843 it shows the following:

    2011-10-18 19:05:53.7067 [PID=19313] Request: [USER#xxxxx] [HOST#4142843] [IP xxx.xxx.xxx.34] client 6.12.26
    2011-10-18 19:05:53.7246 [PID=19313] [send] effective_ncpus 8 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
    2011-10-18 19:05:53.7246 [PID=19313] [send] effective_ngpus 1 max_jobs_on_host_gpu 999999
    2011-10-18 19:05:53.7246 [PID=19313] [send] Not using matchmaker scheduling; Not using EDF sim
    2011-10-18 19:05:53.7246 [PID=19313] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
    2011-10-18 19:05:53.7246 [PID=19313] [send] CUDA: req 1.00 sec, 0.07 instances; est delay 0.00
    2011-10-18 19:05:53.7246 [PID=19313] [send] work_req_seconds: 0.00 secs
    2011-10-18 19:05:53.7246 [PID=19313] [send] available disk 17.88 GB, work_buf_min 0
    2011-10-18 19:05:53.7246 [PID=19313] [send] active_frac 0.999621 on_frac 0.959780 DCF 1.000000
    2011-10-18 19:05:53.7253 [PID=19313] [send] [HOST#4142843] is reliable
    2011-10-18 19:05:53.7253 [PID=19313] [send] set_trust: random choice for error rate 0.013527: no
    2011-10-18 19:05:53.7378 [PID=19313] [version] Don't need CPU jobs, skipping version 23 for hsgamma_FGRP1 ()
    2011-10-18 19:05:53.7378 [PID=19313] [version] no app version available: APP#17 (hsgamma_FGRP1) PLATFORM#2 (windows_intelx86) min_version 0
    2011-10-18 19:05:53.7379 [PID=19313] [version] Checking plan class 'BRP3cuda32'
    2011-10-18 19:05:53.7382 [PID=19313] [version] reading plan classes from file '../plan_class_spec.xml'
    2011-10-18 19:05:53.7382 [PID=19313] [version] driver version required min: 26000, supplied: 0
    2011-10-18 19:05:53.7382 [PID=19313] [version] Checking plan class 'BRP3SSE'
    2011-10-18 19:05:53.7382 [PID=19313] [version] parsed project prefs setting 'also_run_cpu' : true : 1.000000
    2011-10-18 19:05:53.7383 [PID=19313] [version] project prefs setting 'also_run_cpu' (1.000000) prevents using plan class.
    2011-10-18 19:05:53.7383 [PID=19313] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
    2011-10-18 19:05:54.0936 [PID=19313] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket ()
    2011-10-18 19:05:54.0936 [PID=19313] [version] Checking plan class 'SSE2'
    2011-10-18 19:05:54.0936 [PID=19313] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket (SSE2)
    2011-10-18 19:05:54.0936 [PID=19313] [version] no app version available: APP#16 (einstein_S6Bucket) PLATFORM#2 (windows_intelx86) min_version 0
    2011-10-18 19:05:54.0954 [PID=19313] [debug] [HOST#4142843] MSG(high) No work sent
    2011-10-18 19:05:54.0954 [PID=19313] [debug] [HOST#4142843] MSG(high) see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/4142/4142843
    2011-10-18 19:05:54.0954 [PID=19313] Sending reply to [HOST#4142843]: 0 results, delay req 60.00
    2011-10-18 19:05:54.0957 [PID=19313] Scheduler ran 0.396 seconds

    Could someone help
    thanks
    cc


    Bernd Machenschalk
    Bernd Machenschalk
    Moderator
    Administrator
    Joined: 15 Oct 04
    Posts: 4312
    Credit: 250610767
    RAC: 34728

    Can't get any new wu

    18 Oct 2011 19:26:34 UTC
    Message 106979
    (moderation:
      )

      Apparently your Client doesn't report a display driver version ("driver version required min: 26000, supplied: 0"). Anyone knows why? Is this a bug or a feature of 6.12.26, or maybe a problem with the driver?

      BM

      BM

      candido
      candido
      Joined: 8 Mar 11
      Posts: 11
      Credit: 1005637
      RAC: 0

      I cheked the driver, it's

      18 Oct 2011 20:16:23 UTC
      Message 106980
      (moderation:
        )

        I cheked the driver, it's 267.44, clearly inside the requirements for running the wu.
        This is a toshiba laptop and the updates are made through thoshiba, not nvidia.
        I suspect toshiba suplies slightly different versions of the drivers, and that might be the reason the driver version is not being recognise .
        what is surprising is that this computer has been running einstein for some months now without any problems...
        Help appreciated
        cc


        Claggy
        Claggy
        Joined: 29 Dec 06
        Posts: 560
        Credit: 2699403
        RAC: 0

        Looking at your task list

        18 Oct 2011 21:15:56 UTC
        Message 106981 in response to message 106980
        (moderation:
          )

          Looking at your task list shows a few completed Cuda tasks Boinc using 6.12.26:

          Quote:

          6.12.26

          Activated exception handling...
          [23:52:39][9732][INFO ] Starting data processing...
          [23:52:40][9732][INFO ] CUDA global memory status (initial GPU state, including context):
          ------> Used in total: 17 MB (977 MB free / 994 MB total) -> Used by this application (assuming a single GPU task): 0 MB
          [23:52:40][9732][INFO ] Using CUDA device #0 "GeForce GT 540M" (96 CUDA cores / 258.05 GFLOPS)
          [23:52:40][9732][INFO ] Version of installed CUDA driver: 3020
          [23:52:40][9732][INFO ] Version of CUDA driver API used: 3020
          [23:52:40][9732][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory).
          ------> Starting from scratch...

          Have you tried restarting windows since then? and have you changed driver versions since 11 October?

          Claggy

          Bernd Machenschalk
          Bernd Machenschalk
          Moderator
          Administrator
          Joined: 15 Oct 04
          Posts: 4312
          Credit: 250610767
          RAC: 34728

          The server side code has

          18 Oct 2011 21:36:24 UTC
          Message 106982 in response to message 106981
          (moderation:
            )

            The server side code has changed. It appears that the previous code allowed to send CUDA work to clients which didn't report a driver version. This, however, was unintentional.

            To understand how to deal with that I need to know under which conditions a Client doesn't report the driver version.

            BM

            BM

            candido
            candido
            Joined: 8 Mar 11
            Posts: 11
            Credit: 1005637
            RAC: 0

            I have not updated the driver

            18 Oct 2011 21:39:52 UTC
            Message 106983 in response to message 106981
            (moderation:
              )

              I have not updated the driver manually, it might have been an automatic update.
              I am going to restart windows as it has been running for a few days and will get back.


              candido
              candido
              Joined: 8 Mar 11
              Posts: 11
              Credit: 1005637
              RAC: 0

              I have now restarted the

              18 Oct 2011 22:06:22 UTC
              Message 106984
              (moderation:
                )

                I have now restarted the computer and upated the project on the boinc manager but with the same response:

                2011-10-18 21:55:29.2351 [PID=10825] Request: [USER#xxxxx] [HOST#4142843] [IP xxx.xxx.xxx.34] client 6.12.26
                2011-10-18 21:55:29.2365 [PID=10825] [send] effective_ncpus 8 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
                2011-10-18 21:55:29.2365 [PID=10825] [send] effective_ngpus 1 max_jobs_on_host_gpu 999999
                2011-10-18 21:55:29.2365 [PID=10825] [send] Not using matchmaker scheduling; Not using EDF sim
                2011-10-18 21:55:29.2365 [PID=10825] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
                2011-10-18 21:55:29.2365 [PID=10825] [send] CUDA: req 1.00 sec, 0.07 instances; est delay 0.00
                2011-10-18 21:55:29.2365 [PID=10825] [send] work_req_seconds: 0.00 secs
                2011-10-18 21:55:29.2365 [PID=10825] [send] available disk 17.85 GB, work_buf_min 0
                2011-10-18 21:55:29.2365 [PID=10825] [send] active_frac 0.999590 on_frac 0.960138 DCF 0.823854
                2011-10-18 21:55:29.2370 [PID=10825] [send] [HOST#4142843] is reliable
                2011-10-18 21:55:29.2371 [PID=10825] [send] set_trust: random choice for error rate 0.013527: yes
                2011-10-18 21:55:29.2491 [PID=10825] [version] Don't need CPU jobs, skipping version 23 for hsgamma_FGRP1 ()
                2011-10-18 21:55:29.2492 [PID=10825] [version] no app version available: APP#17 (hsgamma_FGRP1) PLATFORM#2 (windows_intelx86) min_version 0
                2011-10-18 21:55:29.2492 [PID=10825] [version] Checking plan class 'BRP3cuda32'
                2011-10-18 21:55:29.2495 [PID=10825] [version] reading plan classes from file '../plan_class_spec.xml'
                2011-10-18 21:55:29.2496 [PID=10825] [version] driver version required min: 26000, supplied: 0
                2011-10-18 21:55:29.2496 [PID=10825] [version] Checking plan class 'BRP3SSE'
                2011-10-18 21:55:29.2496 [PID=10825] [version] parsed project prefs setting 'also_run_cpu' : true : 1.000000
                2011-10-18 21:55:29.2496 [PID=10825] [version] project prefs setting 'also_run_cpu' (1.000000) prevents using plan class.
                2011-10-18 21:55:29.2496 [PID=10825] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
                2011-10-18 21:55:29.5446 [PID=10825] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket ()
                2011-10-18 21:55:29.5446 [PID=10825] [version] Checking plan class 'SSE2'
                2011-10-18 21:55:29.5446 [PID=10825] [version] Don't need CPU jobs, skipping version 101 for einstein_S6Bucket (SSE2)
                2011-10-18 21:55:29.5446 [PID=10825] [version] no app version available: APP#16 (einstein_S6Bucket) PLATFORM#2 (windows_intelx86) min_version 0
                2011-10-18 21:55:29.5466 [PID=10825] [debug] [HOST#4142843] MSG(high) No work sent
                2011-10-18 21:55:29.5466 [PID=10825] [debug] [HOST#4142843] MSG(high) see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/4142/4142843
                2011-10-18 21:55:29.5466 [PID=10825] Sending reply to [HOST#4142843]: 0 results, delay req 60.00
                2011-10-18 21:55:29.5469 [PID=10825] Scheduler ran 0.318 seconds

                At this point I wouldn't know what to do.
                Help is much appreciated.
                Thanks
                cc


                Bernd Machenschalk
                Bernd Machenschalk
                Moderator
                Administrator
                Joined: 15 Oct 04
                Posts: 4312
                Credit: 250610767
                RAC: 34728

                Server configuration has been

                19 Oct 2011 9:22:39 UTC
                Message 106985
                (moderation:
                  )

                  Server configuration has been changed.

                  Please try again.

                  BM

                  BM

                  candido
                  candido
                  Joined: 8 Mar 11
                  Posts: 11
                  Credit: 1005637
                  RAC: 0

                  Many thanks Bernd I will try

                  19 Oct 2011 10:15:07 UTC
                  Message 106986 in response to message 106985
                  (moderation:
                    )

                    Many thanks Bernd
                    I will try latter at night when I return home
                    thanks again
                    cc


                    candido
                    candido
                    Joined: 8 Mar 11
                    Posts: 11
                    Credit: 1005637
                    RAC: 0

                    I have downloaded 6 wu. One

                    19 Oct 2011 18:32:48 UTC
                    Message 106987
                    (moderation:
                      )

                      I have downloaded 6 wu.
                      One almost completed.
                      Thanks
                      cc


                      Richard Haselgrove
                      Richard Haselgrove
                      Joined: 10 Dec 05
                      Posts: 2143
                      Credit: 2959696112
                      RAC: 703689

                      RE: The server side code

                      22 Oct 2011 9:27:30 UTC
                      Message 106988 in response to message 106982
                      (moderation:
                        )

                        Quote:

                        The server side code has changed. It appears that the previous code allowed to send CUDA work to clients which didn't report a driver version. This, however, was unintentional.

                        To understand how to deal with that I need to know under which conditions a Client doesn't report the driver version.

                        BM


                        My laptop host 3868392 is one of the ones which doesn't report a driver version. We used it as a testbed, and found that it's related to having multiple graphics adapters in the host. My laptop has Intel Optimus technology, with a low-power Intel HD graphics chipset for extended battery life when undemanding applications are running, and a NVIDIA GT420M for more demanding work.

                        The current BOINC driver detection logic sees that the first adapter isn't cuda-capable, and doesn't bother checking the others.

                        Our research led to David checking in http://boinc.berkeley.edu/trac/changeset/24046, but as with so much BOINC development, that change is only active in the highly experimental v6.13.xx versions, and hasn't been back-ported in to the bugfix updates of the current recommended line, like the v6.12.41 that I'm running now. You'll have to speak to David or Rom about that.

                        Comment viewing options

                        Select your preferred way to display the comments and click "Save settings" to activate your changes.

                        Forums › Problems and Bug Reports