Many GPU problems

Forums › Problems and Bug Reports

NottiusMaximus
NottiusMaximus
Joined: 4 Jun 05
Posts: 9
Credit: 10366691
RAC: 0
5 Apr 2013 18:38:33 UTC
Topic 196885
(moderation:
    )

    A couple of weeks ago I started receiving errors along the line of GPU missing. Now I'm getting a different error message. App version needs CUDA but GPU doesn't support it.

    Any ideas?

    Here is the first part of my Event Log

    Fri 05 Apr 2013 01:15:23 AM MDT | | Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu
    Fri 05 Apr 2013 01:15:23 AM MDT | | log flags: file_xfer, sched_ops, task
    Fri 05 Apr 2013 01:15:23 AM MDT | | Libraries: libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
    Fri 05 Apr 2013 01:15:23 AM MDT | | Data directory: /var/lib/boinc-client
    Fri 05 Apr 2013 01:15:23 AM MDT | | Processor: 6 AuthenticAMD AMD FX(tm)-6100 Six-Core Processor [Family 21 Model 1 Stepping 2]
    Fri 05 Apr 2013 01:15:23 AM MDT | | Processor: 2.00 MB cache
    Fri 05 Apr 2013 01:15:23 AM MDT | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core arat cpb npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold
    Fri 05 Apr 2013 01:15:23 AM MDT | | OS: Linux: 3.2.0-39-generic
    Fri 05 Apr 2013 01:15:23 AM MDT | | Memory: 7.78 GB physical, 94.00 MB virtual
    Fri 05 Apr 2013 01:15:23 AM MDT | | Disk: 63.80 GB total, 43.89 GB free
    Fri 05 Apr 2013 01:15:23 AM MDT | | Local time is UTC -6 hours
    Fri 05 Apr 2013 01:15:23 AM MDT | | OpenCL: NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 295.40, device version OpenCL 1.1 CUDA, 1024MB, 1024MB available)
    Fri 05 Apr 2013 01:15:23 AM MDT | | App version needs CUDA but GPU doesn't support it
    Fri 05 Apr 2013 01:15:23 AM MDT | Einstein@Home | Application uses missing NVIDIA GPU
    Fri 05 Apr 2013 01:15:23 AM MDT | Einstein@Home | Missing coprocessor for task p2030.20121230.G203.34-01.01.S.b5s0g0.00000_1896_0

    Jord
    Joined: 26 Jan 05
    Posts: 2952
    Credit: 5893653
    RAC: 92

    Many GPU problems

    5 Apr 2013 19:57:48 UTC
    Message 115851
    (moderation:
      )

      It would appear that BOINC isn't able to detect the CUDA part of the drivers. It is capable of seeing that it is OpenCL capable, meaning that BOINC can see the OpenCL library.

      Is this a BOINC version from Berkeley, or repositories?
      And what of the videocard drivers, got those from Nvidia or repositories?

      NottiusMaximus
      NottiusMaximus
      Joined: 4 Jun 05
      Posts: 9
      Credit: 10366691
      RAC: 0

      BOINC is from Kubuntu 12.04

      5 Apr 2013 20:02:09 UTC
      Message 115852 in response to message 115851
      (moderation:
        )

        BOINC is from Kubuntu 12.04 LTS repository. Initially BOINC worked fine, but started to do this just a few weeks ago.

        NottiusMaximus
        NottiusMaximus
        Joined: 4 Jun 05
        Posts: 9
        Credit: 10366691
        RAC: 0

        Video card drivers are

        5 Apr 2013 20:04:23 UTC
        Message 115853 in response to message 115851
        (moderation:
          )

          Video card drivers are NVIDIA-current. I think it's from the Kubuntu reps as well. Not certain tho.

          AgentB
          AgentB
          Joined: 17 Mar 12
          Posts: 915
          Credit: 513211304
          RAC: 0

          RE: Video card drivers are

          6 Apr 2013 0:16:46 UTC
          Message 115854 in response to message 115853
          (moderation:
            )

            Quote:
            Video card drivers are NVIDIA-current. I think it's from the Kubuntu reps as well. Not certain tho.

            Well i´ll hazard a guess...

            The *buntu repositories are always out of date - well at least for 10.04 LTS (¨yes i know it´s April 2013¨, said the pot to the kettle)

            I am running quite happily the nvidia drivers 310.14, i have found the the 304.60 the in the ubuntu reps does not run well here for me.

            If you could point out

            the hostid in question
            if the problem affects all tasks or just occasionally
            and the nvidia driver version,

            that would help.

            NottiusMaximus
            NottiusMaximus
            Joined: 4 Jun 05
            Posts: 9
            Credit: 10366691
            RAC: 0

            I am trying 310 from the

            11 Apr 2013 16:46:51 UTC
            Message 115855 in response to message 115854
            (moderation:
              )

              I am trying 310 from the additional drivers dialog and this is the result. Now BOINC is not recognizing my GPU at all.

              Thu 11 Apr 2013 10:41:27 AM MDT | | Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu
              Thu 11 Apr 2013 10:41:27 AM MDT | | log flags: file_xfer, sched_ops, task
              Thu 11 Apr 2013 10:41:27 AM MDT | | Libraries: libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
              Thu 11 Apr 2013 10:41:27 AM MDT | | Data directory: /var/lib/boinc-client
              Thu 11 Apr 2013 10:41:27 AM MDT | | Processor: 6 AuthenticAMD AMD FX(tm)-6100 Six-Core Processor [Family 21 Model 1 Stepping 2]
              Thu 11 Apr 2013 10:41:27 AM MDT | | Processor: 2.00 MB cache
              Thu 11 Apr 2013 10:41:27 AM MDT | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core arat cpb npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold
              Thu 11 Apr 2013 10:41:27 AM MDT | | OS: Linux: 3.2.0-40-generic
              Thu 11 Apr 2013 10:41:27 AM MDT | | Memory: 7.78 GB physical, 94.00 MB virtual
              Thu 11 Apr 2013 10:41:27 AM MDT | | Disk: 63.80 GB total, 43.13 GB free
              Thu 11 Apr 2013 10:41:27 AM MDT | | Local time is UTC -6 hours
              Thu 11 Apr 2013 10:41:27 AM MDT | | No usable GPUs found
              Thu 11 Apr 2013 10:41:27 AM MDT | Einstein@Home | Application uses missing NVIDIA GPU
              Thu 11 Apr 2013 10:41:27 AM MDT | Einstein@Home | Missing coprocessor for task p2030.20121230.G203.34-01.01.S.b5s0g0.00000_1896_0

              AgentB
              AgentB
              Joined: 17 Mar 12
              Posts: 915
              Credit: 513211304
              RAC: 0

              I may be leading you astray

              11 Apr 2013 20:53:01 UTC
              Message 115856 in response to message 115855
              (moderation:
                )

                I may be leading you astray here as your versions of boinc are later than mine, and my boinc skills are meagre but i had found strange behavior of gpus disappearing at start-up in my early days at Primegrid.

                It might be worth looking at this thread and in particular first try the suggestion relating to boinc-client restart - and good luck!

                NottiusMaximus
                NottiusMaximus
                Joined: 4 Jun 05
                Posts: 9
                Credit: 10366691
                RAC: 0

                I've tried updating

                19 May 2013 16:53:01 UTC
                Message 115857 in response to message 115855
                (moderation:
                  )

                  I've tried updating everything.
                  Now running Kubuntu 13.25 64 bit currently latest greatest.
                  Also running boinc 7.0.65

                  Still not finding my NVIDIA 550TI

                  Here is the first part of my event log

                  Sun 19 May 2013 10:37:13 AM MDT | | Starting BOINC client version 7.0.65 for x86_64-pc-linux-gnu
                  Sun 19 May 2013 10:37:13 AM MDT | | log flags: file_xfer, sched_ops, task
                  Sun 19 May 2013 10:37:13 AM MDT | | Libraries: libcurl/7.29.0 OpenSSL/1.0.1c zlib/1.2.7 libidn/1.25 librtmp/2.3
                  Sun 19 May 2013 10:37:13 AM MDT | | Data directory: /var/lib/boinc-client
                  Sun 19 May 2013 10:37:13 AM MDT | | Processor: 6 AuthenticAMD AMD FX(tm)-6100 Six-Core Processor [Family 21 Model 1 Stepping 2]
                  Sun 19 May 2013 10:37:13 AM MDT | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core arat cpb hw_pstate npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold
                  Sun 19 May 2013 10:37:13 AM MDT | | OS: Linux: 3.8.0-21-generic
                  Sun 19 May 2013 10:37:13 AM MDT | | Memory: 7.78 GB physical, 94.00 MB virtual
                  Sun 19 May 2013 10:37:13 AM MDT | | Disk: 63.68 GB total, 46.84 GB free
                  Sun 19 May 2013 10:37:13 AM MDT | | Local time is UTC -6 hours
                  Sun 19 May 2013 10:37:13 AM MDT | | No usable GPUs found
                  Sun 19 May 2013 10:37:13 AM MDT | | Config: GUI RPCs allowed from:
                  Sun 19 May 2013 10:37:13 AM MDT | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 7046471; resource share 50
                  Sun 19 May 2013 10:37:13 AM MDT | LHC@home 1.0 | URL http://lhcathomeclassic.cern.ch/sixtrack/; Computer ID 10285261; resource share 100
                  Sun 19 May 2013 10:37:13 AM MDT | rosetta@home | URL http://boinc.bakerlab.org/rosetta/; Computer ID 1611531; resource share 50
                  Sun 19 May 2013 10:37:13 AM MDT | | General prefs: from http://setiathome.berkeley.edu/ (last modified 06-Jun-2012 22:25:46)
                  Sun 19 May 2013 10:37:13 AM MDT | | Host location: none
                  Sun 19 May 2013 10:37:13 AM MDT | | General prefs: using your defaults
                  Sun 19 May 2013 10:37:13 AM MDT | | Reading preferences override file
                  Sun 19 May 2013 10:37:13 AM MDT | | Preferences:
                  Sun 19 May 2013 10:37:13 AM MDT | | max memory usage when active: 6374.34MB
                  Sun 19 May 2013 10:37:13 AM MDT | | max memory usage when idle: 7569.53MB
                  Sun 19 May 2013 10:37:13 AM MDT | | max disk usage: 47.12GB
                  Sun 19 May 2013 10:37:13 AM MDT | | max CPUs used: 5
                  Sun 19 May 2013 10:37:13 AM MDT | | suspend work if non-BOINC CPU load exceeds 25 %
                  Sun 19 May 2013 10:37:13 AM MDT | | (to change preferences, visit a project web site or select Preferences in the Manager)
                  Sun 19 May 2013 10:37:13 AM MDT | | gui_rpc_auth.cfg is empty - no GUI RPC password protection
                  Sun 19 May 2013 10:37:13 AM MDT | | Not using a proxy

                  NottiusMaximus
                  NottiusMaximus
                  Joined: 4 Jun 05
                  Posts: 9
                  Credit: 10366691
                  RAC: 0

                  Now I'm just confused as

                  19 Jun 2013 19:07:48 UTC
                  Message 115858 in response to message 115857
                  (moderation:
                    )

                    Now I'm just confused as BOINC. Finds GPU 50% of the time and still won't process.

                    Wed 19 Jun 2013 12:50:20 PM MDT | | Starting BOINC client version 7.0.65 for x86_64-pc-linux-gnu
                    Wed 19 Jun 2013 12:50:20 PM MDT | | log flags: file_xfer, sched_ops, task
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Libraries: libcurl/7.29.0 OpenSSL/1.0.1c zlib/1.2.7 libidn/1.25 librtmp/2.3
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Data directory: /var/lib/boinc-client
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Processor: 6 AuthenticAMD AMD FX(tm)-6100 Six-Core Processor [Family 21 Model 1 Stepping 2]
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core arat cpb hw_pstate npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold
                    Wed 19 Jun 2013 12:50:20 PM MDT | | OS: Linux: 3.8.0-25-generic
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Memory: 7.78 GB physical, 94.00 MB virtual
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Disk: 63.68 GB total, 45.77 GB free
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Local time is UTC -6 hours
                    Wed 19 Jun 2013 12:50:20 PM MDT | | OpenCL: NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 313.30, device version OpenCL 1.1 CUDA, 1023MB, 1023MB available, 116 GFLOPS peak)
                    Wed 19 Jun 2013 12:50:20 PM MDT | | App version needs CUDA but GPU doesn't support it
                    Wed 19 Jun 2013 12:50:20 PM MDT | Einstein@Home | Application uses missing NVIDIA GPU
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Config: GUI RPCs allowed from:
                    Wed 19 Jun 2013 12:50:20 PM MDT | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 7046471; resource share 50
                    Wed 19 Jun 2013 12:50:20 PM MDT | LHC@home 1.0 | URL http://lhcathomeclassic.cern.ch/sixtrack/; Computer ID 10285261; resource share 100
                    Wed 19 Jun 2013 12:50:20 PM MDT | rosetta@home | URL http://boinc.bakerlab.org/rosetta/; Computer ID 1611531; resource share 50
                    Wed 19 Jun 2013 12:50:20 PM MDT | | General prefs: from http://setiathome.berkeley.edu/ (last modified 06-Jun-2012 22:25:46)
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Host location: none
                    Wed 19 Jun 2013 12:50:20 PM MDT | | General prefs: using your defaults
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Reading preferences override file
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Preferences:
                    Wed 19 Jun 2013 12:50:20 PM MDT | | max memory usage when active: 6374.36MB
                    Wed 19 Jun 2013 12:50:20 PM MDT | | max memory usage when idle: 7569.56MB
                    Wed 19 Jun 2013 12:50:20 PM MDT | | max disk usage: 46.20GB
                    Wed 19 Jun 2013 12:50:20 PM MDT | | suspend work if non-BOINC CPU load exceeds 25 %
                    Wed 19 Jun 2013 12:50:20 PM MDT | | (to change preferences, visit a project web site or select Preferences in the Manager)
                    Wed 19 Jun 2013 12:50:20 PM MDT | | gui_rpc_auth.cfg is empty - no GUI RPC password protection
                    Wed 19 Jun 2013 12:50:20 PM MDT | | Not using a proxy
                    Wed 19 Jun 2013 12:50:20 PM MDT | Einstein@Home | Restarting task h1_0501.35_S6GC1__S6BucketLVEa_501.463682292Hz_479_1 using einstein_S6BucketLVE version 104 (SSE2) in slot 2
                    Wed 19 Jun 2013 12:50:20 PM MDT | LHC@home 1.0 | Restarting task wlxscan2_w1cbb__59__s__64.31_59.32__10.5_11__6__52.5_1_sixvf_boinc148099_0 using sixtrack version 44401 (pni) in slot 1
                    Wed 19 Jun 2013 12:50:20 PM MDT | LHC@home 1.0 | Restarting task wlxscan2_w1cbb__38__s__64.31_59.32__6.5_7__6__45_1_sixvf_boinc124413_2 using sixtrack version 44401 (pni) in slot 9

                    Gundolf Jahn
                    Gundolf Jahn
                    Joined: 1 Mar 05
                    Posts: 1079
                    Credit: 341280
                    RAC: 0

                    RE: Now I'm just confused

                    19 Jun 2013 21:04:49 UTC
                    Message 115859 in response to message 115858
                    (moderation:
                      )

                      Quote:

                      Now I'm just confused as BOINC. Finds GPU 50% of the time and still won't process.

                      Wed 19 Jun 2013 12:50:20 PM MDT | | OpenCL: NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 313.30, device version OpenCL 1.1 CUDA, 1023MB, 1023MB available, 116 GFLOPS peak)
                      Wed 19 Jun 2013 12:50:20 PM MDT | | App version needs CUDA but GPU doesn't support it
                      Wed 19 Jun 2013 12:50:20 PM MDT | Einstein@Home | Application uses missing NVIDIA GPU


                      It only finds the OpenCL part of your GPU, not the CUDA one.

                      Check your drivers! (Driver sweeper, clean install)

                      Gruß,
                      Gundolf

                      Computer sind nicht alles im Leben. (Kleiner Scherz)

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

                      RE: RE: Now I'm just

                      19 Jun 2013 21:28:06 UTC
                      Message 115860 in response to message 115859
                      (moderation:
                        )

                        Quote:
                        Quote:

                        Now I'm just confused as BOINC. Finds GPU 50% of the time and still won't process.

                        Wed 19 Jun 2013 12:50:20 PM MDT | | OpenCL: NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 313.30, device version OpenCL 1.1 CUDA, 1023MB, 1023MB available, 116 GFLOPS peak)
                        Wed 19 Jun 2013 12:50:20 PM MDT | | App version needs CUDA but GPU doesn't support it
                        Wed 19 Jun 2013 12:50:20 PM MDT | Einstein@Home | Application uses missing NVIDIA GPU


                        It only finds the OpenCL part of your GPU, not the CUDA one.

                        Check your drivers! (Driver sweeper, clean install)

                        Gruß,
                        Gundolf


                        Do they do a Driversweeper for Linux?

                        Quote:
                        Wed 19 Jun 2013 12:50:20 PM MDT | | Starting BOINC client version 7.0.65 for x86_64-pc-linux-gnu

                        What happens if he restarts the client, does Cuda support show up?

                        There is a thread on the Boinc forum for Linux GPU fixes:

                        Debian/Ubuntu/Mint/Derivatives - GPU recognition fixes

                        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