Einstein@home runs against my own preference

Forums › Problems and Bug Reports

Tuan
Tuan
Joined: 3 Oct 13
Posts: 4
Credit: 30665
RAC: 0
6 Oct 2013 4:26:09 UTC
Topic 197216
(moderation:
    )

    I set up the preference to download works for NVIDIA only, but BOINC keeps getting CPU and ATI works. Yes, there is an ATI card on my Windows box. There are aslo other projects that work perfectly but this one.

    Holmis
    Joined: 4 Jan 05
    Posts: 1118
    Credit: 1055935564
    RAC: 0

    Einstein@home runs against my own preference

    6 Oct 2013 6:06:39 UTC
    Message 118882
    (moderation:
      )

      Hi and welcome to Einstein@home!

      There are 2 computers listed on your account and both have CPU and ATI work, one of them (ID:9044648) also have CUDA work. Are they separate machines or the same? Have you created separate preferences for another venue? Like home, work or school and are the computer set to the right venue?

      In the server logs from the last contact for computer 9044648 at 3 Oct 2013 12:33:08 UTC (looks like the initial contact as the computer was attached to the project at the same time) it was sent the following work as a response to a work request for 1 second for each for the CPU, Nvidia and ATI:
      1 S6 CasA task
      1 BRP4G-opencl-ati task
      1 BRP4G-cuda32-nv301
      So if you had set ATI work to no before that server contact something seems wrong, but this looks like the initial contact when you joined the project and it would be hard to set up project prefs before this contact.

      The latest server logs for computer 9056542 at 6 Oct 2013 4:29:43 UTC tells me that the computer didn't ask for any work but got the following lost tasks resent:
      1 S6 CasA task
      1 BRP5-opencl-ati
      1 BRP4G-opencl-ati
      I think the preferences for what work one wants to receive is ignored when lost tasks are resent.

      Tuan
      Tuan
      Joined: 3 Oct 13
      Posts: 4
      Credit: 30665
      RAC: 0

      I thought the same reason

      16 Oct 2013 9:14:08 UTC
      Message 118883 in response to message 118882
      (moderation:
        )

        I thought the same reason too. Once the contact is done, it can't be reverted so I just removed the project and waited for few days and it is fixed.

        Secondly, I don't know why there are more than one computer registered. I just got a third one.

        About the preference, I only have 1 preference profile which is home across all projects.

        Update: I merged all the computers. Hope it works flawlessly now.

        Tuan
        Tuan
        Joined: 3 Oct 13
        Posts: 4
        Credit: 30665
        RAC: 0

        Nope, it still downloads ati

        16 Oct 2013 10:54:26 UTC
        Message 118884
        (moderation:
          )

          Nope, it still downloads ati and cpug jobs.

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

          RE: Nope, it still

          16 Oct 2013 17:33:33 UTC
          Message 118885 in response to message 118884
          (moderation:
            )

            Quote:
            Nope, it still downloads ati and cpug jobs.


            Please post the Boinc startup from the Event Log, the first 30 to 40 lines will do. (up to the point where Wu's are started)

            Claggy

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

            I test the client operation

            16 Oct 2013 19:20:24 UTC
            Message 118886
            (moderation:
              )

              I test the client operation of the Use CPU/Use Nvidia GPU/Use ATI GPU/Use Intel GPU selections, the first three have worked faultlessly all the way since Boinc 6.10.19 on each venue,
              the client use of the 'Use Intel GPU' selection at present is inoperative on any client, there is a Bug report in for that at present (more than a month now), just waiting for Boinc dev staff to get around to looking at it.

              For an 'Use Nvidia GPU' only request i only request Nvidia (Cuda) GPU work, and only receive Nvidia work
              (the deselected preferences block the client from asking for work for deselected devices):

              Quote:
              2013-10-16 18:23:05.4401 [PID=30526] Request: [USER#xxxxx] [HOST#2542754] [IP xxx.xxx.xxx.224] client 7.2.23
              2013-10-16 18:23:05.4543 [PID=30526] [send] effective_ncpus 2 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
              2013-10-16 18:23:05.4543 [PID=30526] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
              2013-10-16 18:23:05.4543 [PID=30526] [send] Not using matchmaker scheduling; Not using EDF sim
              2013-10-16 18:23:05.4543 [PID=30526] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:23:05.4543 [PID=30526] [send] CUDA: req 5406.87 sec, 1.00 instances; est delay 0.00
              2013-10-16 18:23:05.4543 [PID=30526] [send] ATI: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:23:05.4544 [PID=30526] [send] available disk 98.42 GB, work_buf_min 8640
              2013-10-16 18:23:05.4544 [PID=30526] [send] active_frac 0.996770 on_frac 0.898206 DCF 2.519647
              2013-10-16 18:23:05.4550 [PID=30526] [send] [HOST#2542754] is reliable
              2013-10-16 18:23:05.4550 [PID=30526] [send] set_trust: random choice for error rate 0.000010: yes
              2013-10-16 18:23:05.4550 [PID=30526] [mixed] sending locality work first (0.3784)
              2013-10-16 18:23:09.8021 [PID=30526] [version] Checking plan class 'SSE2'
              2013-10-16 18:23:09.8029 [PID=30526] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
              2013-10-16 18:23:09.8029 [PID=30526] [version] plan class ok
              2013-10-16 18:23:09.8029 [PID=30526] [version] Don't need CPU jobs, skipping version 105 for einstein_S6CasA (SSE2)
              2013-10-16 18:23:09.8029 [PID=30526] [version] no app version available: APP#24 (einstein_S6CasA) PLATFORM#9 (windows_x86_64) min_version 0
              2013-10-16 18:23:09.8029 [PID=30526] [version] no app version available: APP#24 (einstein_S6CasA) PLATFORM#2 (windows_intelx86) min_version 0
              2013-10-16 18:23:09.8032 [PID=30526] [mixed] sending non-locality work second
              2013-10-16 18:23:09.8152 [PID=30526] [version] Checking plan class 'BRP5-opencl-ati'
              2013-10-16 18:23:09.8152 [PID=30526] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:23:09.8153 [PID=30526] [version] Peak flops supplied: 1.792e+12
              2013-10-16 18:23:09.8153 [PID=30526] [version] plan class ok
              2013-10-16 18:23:09.8153 [PID=30526] [version] Don't need ATI jobs, skipping version 139 for einsteinbinary_BRP5 (BRP5-opencl-ati)
              2013-10-16 18:23:09.8153 [PID=30526] [version] Checking plan class 'BRP5-cuda32'
              2013-10-16 18:23:09.8153 [PID=30526] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:23:09.8153 [PID=30526] [version] driver version required max: -29053, supplied: 32641
              2013-10-16 18:23:09.8153 [PID=30526] [version] Checking plan class 'BRP5-cuda32-nv301'
              2013-10-16 18:23:09.8153 [PID=30526] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:23:09.8153 [PID=30526] [version] Peak flops supplied: 1.792e+11
              2013-10-16 18:23:09.8153 [PID=30526] [version] plan class ok
              2013-10-16 18:23:09.8153 [PID=30526] [version] Checking plan class 'BRP5-opencl-ati'
              2013-10-16 18:23:09.8153 [PID=30526] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:23:09.8153 [PID=30526] [version] Peak flops supplied: 1.792e+12
              2013-10-16 18:23:09.8154 [PID=30526] [version] plan class ok
              2013-10-16 18:23:09.8154 [PID=30526] [version] Don't need ATI jobs, skipping version 139 for einsteinbinary_BRP5 (BRP5-opencl-ati)
              2013-10-16 18:23:09.8154 [PID=30526] [version] Best version of app einsteinbinary_BRP5 is ID 479 (43.63 GFLOPS)
              2013-10-16 18:23:09.8168 [PID=30526] [debug] Sorted list of URLs follows [host timezone: UTC+3600]
              2013-10-16 18:23:09.8168 [PID=30526] [debug] zone=+03600 url=http://einstein2.aei.uni-hannover.de
              2013-10-16 18:23:09.8168 [PID=30526] [debug] zone=-21600 url=http://einstein-dl4.phys.uwm.edu
              2013-10-16 18:23:09.8168 [PID=30526] [debug] zone=-21600 url=http://einstein-dl2.phys.uwm.edu
              2013-10-16 18:23:09.8169 [PID=30526] [debug] zone=-28800 url=http://einstein.ligo.caltech.edu
              2013-10-16 18:23:09.8170 [PID=30526] [send] [HOST#2542754] Sending app_version 479 einsteinbinary_BRP5 2 139 BRP5-cuda32-nv301; 43.63 GFLOPS
              2013-10-16 18:23:09.8189 [PID=30526] [send] est. duration for WU 176578323: unscaled 10314.04 scaled 29026.69
              2013-10-16 18:23:09.8189 [PID=30526] [HOST#2542754] Sending [RESULT#405974267 PA0030_00351_346_0] (est. dur. 29026.69 seconds)
              2013-10-16 18:23:09.8195 [PID=30526] [send] don't need more work
              2013-10-16 18:23:09.8206 [PID=30526] Sending reply to [HOST#2542754]: 1 results, delay req 60.00
              2013-10-16 18:23:09.8210 [PID=30526] Scheduler ran 4.388 seconds

              For an 'Use ATI GPU' only request, I only request ATI work and only receive ATI work:

              Quote:
              2013-10-16 18:34:12.1656 [PID=32213] Request: [USER#xxxxx] [HOST#2542754] [IP xxx.xxx.xxx.224] client 7.2.23
              2013-10-16 18:34:12.1796 [PID=32213] [send] effective_ncpus 2 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
              2013-10-16 18:34:12.1796 [PID=32213] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
              2013-10-16 18:34:12.1796 [PID=32213] [send] Not using matchmaker scheduling; Not using EDF sim
              2013-10-16 18:34:12.1796 [PID=32213] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:34:12.1796 [PID=32213] [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:34:12.1796 [PID=32213] [send] ATI: req 8640.00 sec, 1.00 instances; est delay 0.00
              2013-10-16 18:34:12.1796 [PID=32213] [send] work_req_seconds: 0.00 secs
              2013-10-16 18:34:12.1796 [PID=32213] [send] available disk 98.39 GB, work_buf_min 8640
              2013-10-16 18:34:12.1797 [PID=32213] [send] active_frac 0.996772 on_frac 0.898285 DCF 2.519647
              2013-10-16 18:34:12.1806 [PID=32213] [send] [HOST#2542754] is reliable
              2013-10-16 18:34:12.1806 [PID=32213] [send] set_trust: random choice for error rate 0.000010: yes
              2013-10-16 18:34:12.1807 [PID=32213] [mixed] sending non-locality work first (0.9409)
              2013-10-16 18:34:12.1953 [PID=32213] [version] Checking plan class 'BRP5-opencl-ati'
              2013-10-16 18:34:12.1965 [PID=32213] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
              2013-10-16 18:34:12.1965 [PID=32213] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:34:12.1966 [PID=32213] [version] Peak flops supplied: 1.792e+12
              2013-10-16 18:34:12.1966 [PID=32213] [version] plan class ok
              2013-10-16 18:34:12.1966 [PID=32213] [version] Checking plan class 'BRP5-cuda32'
              2013-10-16 18:34:12.1966 [PID=32213] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:34:12.1966 [PID=32213] [version] driver version required max: -29053, supplied: 32641
              2013-10-16 18:34:12.1966 [PID=32213] [version] Checking plan class 'BRP5-cuda32-nv301'
              2013-10-16 18:34:12.1966 [PID=32213] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:34:12.1966 [PID=32213] [version] Peak flops supplied: 1.792e+11
              2013-10-16 18:34:12.1966 [PID=32213] [version] plan class ok
              2013-10-16 18:34:12.1966 [PID=32213] [version] Don't need CUDA jobs, skipping version 139 for einsteinbinary_BRP5 (BRP5-cuda32-nv301)
              2013-10-16 18:34:12.1967 [PID=32213] [version] Checking plan class 'BRP5-opencl-ati'
              2013-10-16 18:34:12.1967 [PID=32213] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
              2013-10-16 18:34:12.1967 [PID=32213] [version] Peak flops supplied: 1.792e+12
              2013-10-16 18:34:12.1967 [PID=32213] [version] plan class ok
              2013-10-16 18:34:12.1967 [PID=32213] [version] Best version of app einsteinbinary_BRP5 is ID 481 (65.44 GFLOPS)
              2013-10-16 18:34:12.1982 [PID=32213] [debug] Sorted list of URLs follows [host timezone: UTC+3600]
              2013-10-16 18:34:12.1982 [PID=32213] [debug] zone=+03600 url=http://einstein2.aei.uni-hannover.de
              2013-10-16 18:34:12.1982 [PID=32213] [debug] zone=-21600 url=http://einstein-dl4.phys.uwm.edu
              2013-10-16 18:34:12.1982 [PID=32213] [debug] zone=-21600 url=http://einstein-dl2.phys.uwm.edu
              2013-10-16 18:34:12.1982 [PID=32213] [debug] zone=-28800 url=http://einstein.ligo.caltech.edu
              2013-10-16 18:34:12.1985 [PID=32213] [send] [HOST#2542754] Sending app_version 481 einsteinbinary_BRP5 9 139 BRP5-opencl-ati; 65.44 GFLOPS
              2013-10-16 18:34:12.2021 [PID=32213] [send] est. duration for WU 176598755: unscaled 6876.03 scaled 19349.39
              2013-10-16 18:34:12.2021 [PID=32213] [HOST#2542754] Sending [RESULT#406018409 PA0030_00461_122_0] (est. dur. 19349.39 seconds)
              2013-10-16 18:34:12.2032 [PID=32213] [send] don't need more work
              2013-10-16 18:34:12.2032 [PID=32213] [mixed] sending locality work second
              2013-10-16 18:34:12.2034 [PID=32213] [send] don't need more work
              2013-10-16 18:34:12.2034 [PID=32213] [send] don't need more work
              2013-10-16 18:34:12.2046 [PID=32213] Sending reply to [HOST#2542754]: 1 results, delay req 60.00
              2013-10-16 18:34:12.2050 [PID=32213] Scheduler ran 0.047 seconds

              For a 'Use CPU' only request I only request CPU work and only receive CPU work:

              Quote:

              2013-10-16 18:45:39.7582 [PID=1567] Request: [USER#xxxxx] [HOST#2542754] [IP xxx.xxx.xxx.224] client 7.2.23
              2013-10-16 18:45:39.7702 [PID=1567 ] [send] effective_ncpus 2 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] Not using matchmaker scheduling; Not using EDF sim
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] CPU: req 17280.00 sec, 2.00 instances; est delay 0.00
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] ATI: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 18:45:39.7703 [PID=1567 ] [send] work_req_seconds: 17282013-10-16 18:44:43.6553 [PID=1388 ] SCHEDULER_REQUEST::parse(): unrecognized: 0

              2013-10-16 18:45:39.7703 [PID=1567 ] [send] available disk 98.39 GB, work_buf_min 8640
              2013-10-16 18:45:39.7704 [PID=1567 ] [send] active_frac 0.996775 on_frac 0.898365 DCF 2.519647
              2013-10-16 18:45:39.7712 [PID=1567 ] [send] [HOST#2542754] is reliable
              2013-10-16 18:45:39.7713 [PID=1567 ] [send] set_trust: random choice for error rate 0.000010: yes
              2013-10-16 18:45:39.7713 [PID=1567 ] [mixed] sending locality work first (0.1562)
              2013-10-16 18:45:47.9558 [PID=1567 ] [version] Checking plan class 'SSE2'
              2013-10-16 18:45:47.9566 [PID=1567 ] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
              2013-10-16 18:45:47.9567 [PID=1567 ] [version] plan class ok
              2013-10-16 18:45:47.9567 [PID=1567 ] [version] Best version of app einstein_S6CasA is ID 468 (6.98 GFLOPS)
              2013-10-16 18:45:47.9575 [PID=1567 ] [debug] Sorted list of URLs follows [host timezone: UTC+3600]
              2013-10-16 18:45:47.9575 [PID=1567 ] [debug] zone=+03600 url=http://einstein2.aei.uni-hannover.de
              2013-10-16 18:45:47.9575 [PID=1567 ] [debug] zone=-21600 url=http://einstein-dl4.phys.uwm.edu
              2013-10-16 18:45:47.9575 [PID=1567 ] [debug] zone=-21600 url=http://einstein-dl2.phys.uwm.edu
              2013-10-16 18:45:47.9575 [PID=1567 ] [debug] zone=-28800 url=http://einstein.ligo.caltech.edu
              2013-10-16 18:45:47.9576 [PID=1567 ] [send] [HOST#2542754] Sending app_version 468 einstein_S6CasA 2 105 SSE2; 6.98 GFLOPS
              2013-10-16 18:45:47.9585 [PID=1567 ] [send] est. duration for WU 176613065: unscaled 14439.65 scaled 40629.97
              2013-10-16 18:45:47.9585 [PID=1567 ] [HOST#2542754] Sending [RESULT#406050511 h1_0529.90_S6Directed__S6CasAf40a_530.6Hz_9_1] (est. dur. 40629.97 seconds)
              2013-10-16 18:45:47.9629 [PID=1567 ] [send] est. duration for WU 176613068: unscaled 14439.65 scaled 40629.97
              2013-10-16 18:45:47.9629 [PID=1567 ] [send] [WU#176613068] meets deadline: 40629.97 + 40629.97 < 1209600
              2013-10-16 18:45:47.9633 [PID=1567 ] [send] [HOST#2542754] Sending app_version 468 einstein_S6CasA 2 105 SSE2; 6.98 GFLOPS
              2013-10-16 18:45:47.9640 [PID=1567 ] [send] est. duration for WU 176613068: unscaled 14439.65 scaled 40629.97
              2013-10-16 18:45:47.9640 [PID=1567 ] [HOST#2542754] Sending [RESULT#406050518 h1_0529.90_S6Directed__S6CasAf40a_530.6Hz_8_1] (est. dur. 40629.97 seconds)
              2013-10-16 18:45:47.9652 [PID=1567 ] [send] don't need more work
              2013-10-16 18:45:47.9652 [PID=1567 ] [send] don't need more work
              2013-10-16 18:45:47.9652 [PID=1567 ] [mixed] sending non-locality work second
              2013-10-16 18:45:47.9653 [PID=1567 ] [send] don't need more work
              2013-10-16 18:45:47.9666 [PID=1567 ] Sending reply to [HOST#2542754]: 2 results, delay req 60.00
              2013-10-16 18:45:47.9671 [PID=1567 ] Scheduler ran 8.216 seconds

              and if I have all the Use selections deselected Boinc can't ask for work for any device and can't receive it either.
              (the exception is the Einstein server will resend lost work even if work hasn't been requested, other projects with newer Boinc server software won't resent work if it hasn't been asked for):

              Quote:
              2013-10-16 19:05:06.7808 [PID=4830] Request: [USER#xxxxx] [HOST#2542754] [IP xxx.xxx.xxx.224] client 7.2.23
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] effective_ncpus 2 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] Not using matchmaker scheduling; Not using EDF sim
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] ATI: req 0.00 sec, 0.00 instances; est delay 0.00
              2013-10-16 19:05:06.7934 [PID=4830 ] [send] work_req_seconds: 0.00 secs
              2013-10-16 19:05:06.7935 [PID=4830 ] [send] available disk 98.07 GB, work_buf_min 43200
              2013-10-16 19:05:06.7935 [PID=4830 ] [send] active_frac 0.996779 on_frac 0.898502 DCF 2.519647
              2013-10-16 19:05:06.7961 [PID=4830 ] Sending reply to [HOST#2542754]: 0 results, delay req 60.00
              2013-10-16 19:05:06.7964 [PID=4830 ] Scheduler ran 0.023 seconds

              Claggy

              Tuan
              Tuan
              Joined: 3 Oct 13
              Posts: 4
              Credit: 30665
              RAC: 0

              Thanks for your help, I

              17 Oct 2013 23:30:57 UTC
              Message 118887 in response to message 118886
              (moderation:
                )

                Thanks for your help, I aborted those cpu/ati jobs and it works normally now. Btw, I am completely new to the whole BOINC. When you say Event Log, you mean Advanced/Event Log (Ctrl+Shift+E) right?

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

                RE: Thanks for your help, I

                17 Oct 2013 23:44:33 UTC
                Message 118888 in response to message 118887
                (moderation:
                  )

                  Quote:
                  Thanks for your help, I aborted those cpu/ati jobs and it works normally now. Btw, I am completely new to the whole BOINC. When you say Event Log, you mean Advanced/Event Log (Ctrl+Shift+E) right?


                  Yes.

                  Claggy

                  Comment viewing options

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

                  Forums › Problems and Bug Reports