BRP4 1.28 OpenCL and CUDA app : feedback thread

Forums › Problems and Bug Reports

Horacio
Horacio
Joined: 3 Oct 11
Posts: 205
Credit: 80557243
RAC: 0

One thing that seems to be

10 Sep 2012 19:55:13 UTC
Message 111068
(moderation:
    )

    One thing that seems to be missing is an update of the "Estimated app speed" supplied by the servers for the new app...

    In my case the BRP 1.28 on the 560Ti host leads to a DCF of 0.8 while the S6LV1 leads to a DCF of 1.7... I know that the speeds are some kind of predefined averages but AFAIK is still using the speed of the previous version...
    With the DCF varying so much BOINC asks for more CPU work every time I finish a BRP task causing (I guess) unneeded extra load on the schedullers/download servers...

    William J. Watkins
    William J. Watkins
    Joined: 24 Feb 05
    Posts: 4
    Credit: 679522
    RAC: 0

    10/10/2012 7:50:11 AM | |

    11 Oct 2012 2:06:26 UTC
    Message 111069
    (moderation:
      )

      10/10/2012 7:50:11 AM | | Starting BOINC client version 7.0.28 for windows_x86_64
      10/10/2012 7:50:11 AM | | log flags: file_xfer, sched_ops, task
      10/10/2012 7:50:11 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
      10/10/2012 7:50:11 AM | | Data directory: C:\ProgramData\BOINC
      10/10/2012 7:50:11 AM | | Running under account Will
      10/10/2012 7:50:11 AM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz [Family 6 Model 42 Stepping 7]
      10/10/2012 7:50:11 AM | | Processor: 256.00 KB cache
      10/10/2012 7:50:11 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx tm2 popcnt aes pbe
      10/10/2012 7:50:11 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
      10/10/2012 7:50:11 AM | | Memory: 7.95 GB physical, 15.90 GB virtual
      10/10/2012 7:50:11 AM | | Disk: 195.31 GB total, 144.50 GB free
      10/10/2012 7:50:11 AM | | Local time is UTC -6 hours
      10/10/2012 7:50:11 AM | | NVIDIA GPU 0: GeForce GTX 260 (driver version 306.23, CUDA version 5.0, compute capability 1.3, 896MB, 8381209MB available, 805 GFLOPS peak)
      10/10/2012 7:50:11 AM | | OpenCL: NVIDIA GPU 0: GeForce GTX 260 (driver version 306.23, device version OpenCL 1.0 CUDA, 896MB, 8381209MB available)
      10/10/2012 7:50:11 AM | | Config: don't compute while WorldOfTanks.exe is running
      10/10/2012 7:50:11 AM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 5848606; resource share 100
      10/10/2012 7:50:11 AM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6777460; resource share 100
      10/10/2012 7:50:11 AM | | General prefs: from http://www.worldcommunitygrid.org/ (last modified 15-May-2012 11:07:16)
      10/10/2012 7:50:11 AM | | Host location: none
      10/10/2012 7:50:11 AM | | General prefs: using your defaults
      10/10/2012 7:50:11 AM | | Reading preferences override file
      10/10/2012 7:50:11 AM | | Preferences:
      10/10/2012 7:50:11 AM | | max memory usage when active: 4070.55MB
      10/10/2012 7:50:11 AM | | max memory usage when idle: 7327.00MB
      10/10/2012 7:50:11 AM | | max disk usage: 15.00GB
      10/10/2012 7:50:11 AM | | max CPUs used: 2
      10/10/2012 7:50:11 AM | | don't use GPU while active
      10/10/2012 7:50:11 AM | | suspend work if non-BOINC CPU load exceeds 25 %
      10/10/2012 7:50:11 AM | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
      10/10/2012 7:50:11 AM | | Not using a proxy

      Gettin alot of computation errors on the BRP4 1.28 packets. Any ideas?

      I also see alot of this in my logs... :
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_0 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_1 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_2 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_3 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_4 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_5 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_6 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_7 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent
      10/10/2012 9:16:44 AM | Einstein@Home | Starting task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 using einsteinbinary_BRP4 version 128 (BRP4cuda32nv301) in slot 3
      10/10/2012 9:16:45 AM | Einstein@Home | Computation for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 finished
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_0 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_1 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_2 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_3 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_4 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_5 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_6 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0_7 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_224_0 absent
      10/10/2012 9:16:45 AM | Einstein@Home | Starting task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 using einsteinbinary_BRP4 version 128 (BRP4cuda32nv301) in slot 3
      10/10/2012 9:16:47 AM | Einstein@Home | Computation for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 finished
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_0 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_1 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_2 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_3 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_4 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_5 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_6 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0_7 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_928_0 absent
      10/10/2012 9:16:47 AM | Einstein@Home | Starting task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 using einsteinbinary_BRP4 version 128 (BRP4cuda32nv301) in slot 3
      10/10/2012 9:16:48 AM | Einstein@Home | Computation for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 finished
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_0 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_1 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_2 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_3 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_4 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_5 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_6 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0_7 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_352_0 absent
      10/10/2012 9:16:48 AM | Einstein@Home | Starting task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 using einsteinbinary_BRP4 version 128 (BRP4cuda32nv301) in slot 3
      10/10/2012 9:16:50 AM | Einstein@Home | Computation for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 finished
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_0 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_1 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_2 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_3 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_4 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_5 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_6 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent
      10/10/2012 9:16:50 AM | Einstein@Home | Output file p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1_7 for task p2030.20111001.G196.46-00.82.C.b5s0g0.00000_216_1 absent

      William J. Watkins
      William J. Watkins
      Joined: 24 Feb 05
      Posts: 4
      Credit: 679522
      RAC: 0

      This happens immediately when

      11 Oct 2012 2:13:38 UTC
      Message 111070 in response to message 111069
      (moderation:
        )

        This happens immediately when boinc tries to process those tasks.

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

        RE: 10/10/2012 7:50:11 AM |

        12 Oct 2012 20:48:35 UTC
        Message 111071 in response to message 111069
        (moderation:
          )

          Quote:

          10/10/2012 7:50:11 AM | | Starting BOINC client version 7.0.28 for windows_x86_64
          10/10/2012 7:50:11 AM | | log flags: file_xfer, sched_ops, task
          10/10/2012 7:50:11 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
          10/10/2012 7:50:11 AM | | Data directory: C:\ProgramData\BOINC
          10/10/2012 7:50:11 AM | | Running under account Will
          10/10/2012 7:50:11 AM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz [Family 6 Model 42 Stepping 7]
          10/10/2012 7:50:11 AM | | Processor: 256.00 KB cache
          10/10/2012 7:50:11 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx tm2 popcnt aes pbe
          10/10/2012 7:50:11 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
          10/10/2012 7:50:11 AM | | Memory: 7.95 GB physical, 15.90 GB virtual
          10/10/2012 7:50:11 AM | | Disk: 195.31 GB total, 144.50 GB free
          10/10/2012 7:50:11 AM | | Local time is UTC -6 hours
          10/10/2012 7:50:11 AM | | NVIDIA GPU 0: GeForce GTX 260 (driver version 306.23, CUDA version 5.0, compute capability 1.3, 896MB, 8381209MB available, 805 GFLOPS peak)
          10/10/2012 7:50:11 AM | | OpenCL: NVIDIA GPU 0: GeForce GTX 260 (driver version 306.23, device version OpenCL 1.0 CUDA, 896MB, 8381209MB available)
          10/10/2012 7:50:11 AM | | Config: don't compute while WorldOfTanks.exe is running
          10/10/2012 7:50:11 AM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 5848606; resource share 100

          Gettin alot of computation errors on the BRP4 1.28 packets. Any ideas?

          I also see alot of this in my logs... :
          10/10/2012 9:16:44 AM | Einstein@Home | Output file p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1_0 for task p2030.20111001.G196.58-00.59.C.b3s0g0.00000_760_1 absent

          Hi William

          I'll start this by saying that as long as you got you computers hidden anyone who would like to help you with this problem has to work a bit extra for the diagnosis, but thanks to the log you posted I could take the hostid and locate the records for that host anyway.

          Here is a link to host 5848606 for anyone else that's interested

          To get the real error for these task go to you account, click on tasks and the click on a task name.
          Here is the error from one of your tasks:

          7.0.28

          - exit code -1073741515 (0xc0000135)

          ]]>

          I suggest reading through this thread over in the Cruncher's Corner section of the forum.

          William J. Watkins
          William J. Watkins
          Joined: 24 Feb 05
          Posts: 4
          Credit: 679522
          RAC: 0

          10/17/2012 10:52:17 AM | |

          17 Oct 2012 16:55:26 UTC
          Message 111072 in response to message 111071
          (moderation:
            )

            10/17/2012 10:52:17 AM | | Starting BOINC client version 6.12.34 for windows_x86_64
            10/17/2012 10:52:17 AM | | Config: don't compute while WorldOfTanks.exe is running
            10/17/2012 10:52:17 AM | | log flags: file_xfer, sched_ops, task
            10/17/2012 10:52:17 AM | | Libraries: libcurl/7.21.6 OpenSSL/1.0.0d zlib/1.2.5
            10/17/2012 10:52:17 AM | | Data directory: C:\ProgramData\BOINC
            10/17/2012 10:52:17 AM | | Running under account Will
            10/17/2012 10:52:17 AM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz [Family 6 Model 42 Stepping 7]
            10/17/2012 10:52:17 AM | | Processor: 256.00 KB cache
            10/17/2012 10:52:17 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx tm2 popcnt aes pbe
            10/17/2012 10:52:17 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
            10/17/2012 10:52:17 AM | | Memory: 7.51 GB physical, 15.02 GB virtual
            10/17/2012 10:52:17 AM | | Disk: 195.31 GB total, 146.72 GB free
            10/17/2012 10:52:17 AM | | Local time is UTC -6 hours
            10/17/2012 10:52:17 AM | | NVIDIA GPU 0: GeForce GTX 260 (driver version 30623, CUDA version 5000, compute capability 1.3, 896MB, 537 GFLOPS peak)
            10/17/2012 10:52:17 AM | | A new version of BOINC is available. Download it.
            10/17/2012 10:52:17 AM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 5848606; resource share 100
            10/17/2012 10:52:17 AM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6777460; resource share 100
            10/17/2012 10:52:17 AM | | General prefs: from http://www.worldcommunitygrid.org/ (last modified 15-May-2012 11:07:16)
            10/17/2012 10:52:17 AM | | Host location: none
            10/17/2012 10:52:17 AM | | General prefs: using your defaults
            10/17/2012 10:52:17 AM | | Reading preferences override file
            10/17/2012 10:52:17 AM | | Preferences:
            10/17/2012 10:52:17 AM | | max memory usage when active: 3846.55MB
            10/17/2012 10:52:17 AM | | max memory usage when idle: 6923.80MB
            10/17/2012 10:52:17 AM | | max disk usage: 15.00GB
            10/17/2012 10:52:17 AM | | max CPUs used: 2
            10/17/2012 10:52:17 AM | | don't use GPU while active
            10/17/2012 10:52:17 AM | | suspend work if non-BOINC CPU load exceeds 25 %
            10/17/2012 10:52:17 AM | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
            10/17/2012 10:52:17 AM | | Not using a proxy

            Works fine once I downgraded. Even though GPU processing performance went down to the 500Gflops range...

            Horacio
            Horacio
            Joined: 3 Oct 11
            Posts: 205
            Credit: 80557243
            RAC: 0

            RE: Works fine once I

            18 Oct 2012 18:45:57 UTC
            Message 111073 in response to message 111072
            (moderation:
              )

              Quote:
              Works fine once I downgraded. Even though GPU processing performance went down to the 500Gflops range...

              AFAIK, that value is not a measured benchmark. BOINC gets that value doing some maths based on the model of the GPU and other things...
              I guess that, the difference is just because the old version use a different formula, probably because your GPU was not available when that version of BOINC was released... Or because htey find it was wrong and improoved them in new versions... or something alike...

              Richard Haselgrove
              Richard Haselgrove
              Joined: 10 Dec 05
              Posts: 2143
              Credit: 2955093239
              RAC: 714643

              Request for

              19 Oct 2012 11:44:48 UTC
              Message 111074
              (moderation:
                )

                Request for help/information:

                Does anybody here have any experience of running the opencl-ati application (any version) under a Windows ATI driver in the range 11.7-11.11 inclusive?

                Those are the drivers with APP OpenCL runtime support - misleadingly called 'SDK' by AMD - version 2.5

                I know those drivers are unpopular with AMD users, because that runtime uses the CPU inefficiently, but they still crop up sometimes. I'm getting reports from another project that not only are they inefficient, they can also cause science errors in an application running under BOINC. Could one of the mods/scientists tell if there are any indications of invalidated tasks in the database like that?

                William J. Watkins
                William J. Watkins
                Joined: 24 Feb 05
                Posts: 4
                Credit: 679522
                RAC: 0

                Thank you

                19 Oct 2012 23:50:43 UTC
                Message 111075 in response to message 111073
                (moderation:
                  )

                  Thank you

                  Oliver Behnke
                  Oliver Behnke
                  Moderator
                  Administrator
                  Joined: 4 Sep 07
                  Posts: 984
                  Credit: 25171376
                  RAC: 43

                  Hi! RE: I'm getting

                  6 Nov 2012 14:38:37 UTC
                  Message 111076 in response to message 111074
                  (moderation:
                    )

                    Hi!

                    Quote:
                    I'm getting reports from another project that not only are they inefficient, they can also cause science errors in an application running under BOINC. Could one of the mods/scientists tell if there are any indications of invalidated tasks in the database like that?

                    I had a look at our DB and we currently don't have anyone (yep, zero that is) using those driver versions crunching OpenCL tasks on a Microsoft OS.

                    The only OpenCL numerical stability issue I can remember right now are the native_sin/native_cos functions causing a lot of invalid results on 69xx Radeons for instance. These functions are hardware-dependent and the 69xx series seems to be very "bad" in this regard. Beware, Apple's FFT sample does use these functions...

                    Sorry,
                    Oliver

                    Einstein@Home Project

                    Raistmer*
                    Raistmer*
                    Joined: 20 Feb 05
                    Posts: 208
                    Credit: 181403999
                    RAC: 8764

                    And what about Catalyst 12.10

                    7 Nov 2012 6:34:13 UTC
                    Message 111077 in response to message 111076
                    (moderation:
                      )

                      And what about Catalyst 12.10 + HD7xxx combos ?
                      Do you have such in DB? What about their error rate ?

                      Comment viewing options

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

                      Forums › Problems and Bug Reports