ATI issue

Gary Charpentier
Gary Charpentier
Joined: 13 Jun 06
Posts: 1,963
Credit: 99,625,426
RAC: 20,304
Topic 196660

Quote:
11/28/2012 6:16:32 AM | | No config file found - using defaults
11/28/2012 6:16:32 AM | | Starting BOINC client version 7.0.28 for windows_x86_64
11/28/2012 6:16:32 AM | | log flags: file_xfer, sched_ops, task
11/28/2012 6:16:32 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
11/28/2012 6:16:32 AM | | Data directory: C:\ProgramData\BOINC
11/28/2012 6:16:32 AM | | Running under account Odd Fellows Cemetery
11/28/2012 6:16:32 AM | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
11/28/2012 6:16:32 AM | | Processor: 256.00 KB cache
11/28/2012 6:16:32 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 smx tm2 popcnt aes pbe
11/28/2012 6:16:32 AM | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
11/28/2012 6:16:32 AM | | Memory: 9.96 GB physical, 19.91 GB virtual
11/28/2012 6:16:32 AM | | Disk: 284.42 GB total, 222.98 GB free
11/28/2012 6:16:32 AM | | Local time is UTC -8 hours
11/28/2012 6:16:32 AM | | ATI GPU 0: Cedar (CAL version 1.4.1741, 512MB, 479MB available, 208 GFLOPS peak)
11/28/2012 6:16:32 AM | | OpenCL: ATI GPU 0: Cedar (driver version 1016.4 (VM), device version OpenCL 1.2 AMD-APP (1016.4), 512MB, 479MB available)


http://einsteinathome.org/task/320451287
http://einsteinathome.org/task/320451088
and lots more ...

Quote:

7.0.28

Maximum elapsed time exceeded

Activated exception handling...
[04:59:49][3052][INFO ] Starting data processing...
[04:59:51][3052][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc.
[04:59:51][3052][INFO ] Using OpenCL device "Cedar" by: Advanced Micro Devices, Inc.
[04:59:52][3052][WARN ] Kernel "kernelTimeSeriesMeanReduction" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelPowerSpectrum" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelPowerSpectrum_radix3_r2c" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelHarmonicSumming" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items

The maximum time exceeded seems to be the error as nothing else pops out at me in the output. That leaves me wondering if the GPU is so underpowered if it shouldn't be assigned tasks, or is this an issue with the DCF being out of whack, which can't be corrected until a workunit finishes?

Suggestions ...

Presently I've disabled GPU fetch for the machine.

Fred J. Verster
Fred J. Verster
Joined: 27 Apr 08
Posts: 118
Credit: 22,451,438
RAC: 0

ATI issue

Quote:
Quote:
11/28/2012 6:16:32 AM | | No config file found - using defaults
11/28/2012 6:16:32 AM | | Starting BOINC client version 7.0.28 for windows_x86_64
11/28/2012 6:16:32 AM | | log flags: file_xfer, sched_ops, task
11/28/2012 6:16:32 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
11/28/2012 6:16:32 AM | | Data directory: C:\ProgramData\BOINC
11/28/2012 6:16:32 AM | | Running under account Odd Fellows Cemetery
11/28/2012 6:16:32 AM | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
11/28/2012 6:16:32 AM | | Processor: 256.00 KB cache
11/28/2012 6:16:32 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 smx tm2 popcnt aes pbe
11/28/2012 6:16:32 AM | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
11/28/2012 6:16:32 AM | | Memory: 9.96 GB physical, 19.91 GB virtual
11/28/2012 6:16:32 AM | | Disk: 284.42 GB total, 222.98 GB free
11/28/2012 6:16:32 AM | | Local time is UTC -8 hours
11/28/2012 6:16:32 AM | | ATI GPU 0: Cedar (CAL version 1.4.1741, 512MB, 479MB available, 208 GFLOPS peak)
11/28/2012 6:16:32 AM | | OpenCL: ATI GPU 0: Cedar (driver version 1016.4 (VM), device version OpenCL 1.2 AMD-APP (1016.4), 512MB, 479MB available)

http://einsteinathome.org/task/320451287
http://einsteinathome.org/task/320451088
and lots more ...
Quote:

7.0.28

Maximum elapsed time exceeded

Activated exception handling...
[04:59:49][3052][INFO ] Starting data processing...
[04:59:51][3052][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc.
[04:59:51][3052][INFO ] Using OpenCL device "Cedar" by: Advanced Micro Devices, Inc.
[04:59:52][3052][WARN ] Kernel "kernelTimeSeriesMeanReduction" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelPowerSpectrum" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelPowerSpectrum_radix3_r2c" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items
[04:59:52][3052][WARN ] Kernel "kernelHarmonicSumming" exceeds device-specific maximum work group size (requested: 256)!
------> Reducing kernel's work group size to allowed maximum of: 128 work items

The maximum time exceeded seems to be the error as nothing else pops out at me in the output. That leaves me wondering if the GPU is so underpowered if it shouldn't be assigned tasks, or is this an issue with the DCF being out of whack, which can't be corrected until a workunit finishes?

Suggestions ...

Presently I've disabled GPU fetch for the machine.

The culprit is the Work Group Size, which could be changed
through a cmd-line option, in an app_info.xml file but Einstein uses another way these days. Also the # tasks per GPU can be changed, besides other options.
If I'm not teribly mistaken.

I don't know if this is possible at Einstein?!

Hope someone with some solution will chime in.
(I use the CYPRESS (HD5870) GPUs (2), which has a 256 Work Group Size).

Comment viewing options

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