NVidea GPU doesn't support openCL message; missing GPU

Message boards : Number crunching : NVidea GPU doesn't support openCL message; missing GPU
Message board moderation

To post messages, you must log in.

AuthorMessage
SNicol

Send message
Joined: 26 Jun 99
Posts: 4
Credit: 140,565,799
RAC: 52
United States
Message 1339480 - Posted: 19 Feb 2013, 23:15:05 UTC


In a recently built system running Win 7 Ultimate, 64-bit. I decided to rebuild the system on Windows 8. It was horrible, so I reinstalled Win 7.
Since then it has not been downloading GPU work, or if it has,
it has been mostly non-existent. This hardware worked beautifully with the GPUs previously. It says non-existent GPU. What have I missed?
BOINC 7.0.28 is what the web installer recommends for this OS. The hardware detected appears correct.
I have also set the CUDA_GRID_SIZE_COMPAT environment variable to 1.


Here is the log file after repairing BOINC and restarting the machine.


2/19/2013 5:25:00 PM | | Starting BOINC client version 7.0.28 for windows_x86_64
2/19/2013 5:25:00 PM | | log flags: file_xfer, sched_ops, task
2/19/2013 5:25:00 PM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
2/19/2013 5:25:00 PM | | Data directory: C:\ProgramData\BOINC
2/19/2013 5:25:00 PM | | Running under account snicol
2/19/2013 5:25:00 PM | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz [Family 6 Model 58 Stepping 9]
2/19/2013 5:25:00 PM | | Processor: 256.00 KB cache
2/19/2013 5:25:00 PM | | 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
2/19/2013 5:25:00 PM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
2/19/2013 5:25:00 PM | | Memory: 15.89 GB physical, 31.77 GB virtual
2/19/2013 5:25:00 PM | | Disk: 223.47 GB total, 161.61 GB free
2/19/2013 5:25:00 PM | | Local time is UTC -5 hours
2/19/2013 5:25:00 PM | | NVIDIA GPU 0: GeForce GTX 680 (driver version 310.90, CUDA version 5.0, compute capability 3.0, 2048MB, 8382294MB available, 3453 GFLOPS peak)
2/19/2013 5:25:00 PM | | NVIDIA GPU 1: GeForce GTX 680 (driver version 310.90, CUDA version 5.0, compute capability 3.0, 2048MB, 1887MB available, 3453 GFLOPS peak)
2/19/2013 5:25:00 PM | | OpenCL: NVIDIA GPU 0 (not used): GeForce GTX 680 (driver version 310.90, device version OpenCL 1.1 CUDA, 2048MB, 8382294MB available)
2/19/2013 5:25:00 PM | | OpenCL: NVIDIA GPU 1 (not used): GeForce GTX 680 (driver version 310.90, device version OpenCL 1.1 CUDA, 2048MB, 1887MB available)
2/19/2013 5:25:00 PM | | App version needs OpenCL but GPU doesn't support it
2/19/2013 5:25:00 PM | SETI@home | Application uses missing NVIDIA GPU
2/19/2013 5:25:00 PM | | Config: use all coprocessors
2/19/2013 5:25:00 PM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6893033; resource share 100
2/19/2013 5:25:00 PM | SETI@home | General prefs: from SETI@home (last modified 21-Oct-2012 22:20:55)
2/19/2013 5:25:00 PM | SETI@home | Computer location: home
2/19/2013 5:25:00 PM | | General prefs: using separate prefs for home
2/19/2013 5:25:00 PM | | Reading preferences override file
2/19/2013 5:25:00 PM | | Preferences:
2/19/2013 5:25:00 PM | | max memory usage when active: 8133.18MB
2/19/2013 5:25:00 PM | | max memory usage when idle: 14639.72MB
2/19/2013 5:25:00 PM | | max disk usage: 10.00GB
2/19/2013 5:25:00 PM | | don't compute while active
2/19/2013 5:25:00 PM | | don't use GPU while active
2/19/2013 5:25:00 PM | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
2/19/2013 5:25:00 PM | | Not using a proxy
2/19/2013 5:25:01 PM | | Suspending computation - computer is in use


Here is the content of cc_config.xml found in C:/ProgramData/BOINC/

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
</cc_config>

Thanks for any advice. I did not recognize this issue being explained for Nvidea GPUs in the first few pages of threads.

Steve
ID: 1339480 · Report as offensive
Profile Gundolf Jahn

Send message
Joined: 19 Sep 00
Posts: 3184
Credit: 446,358
RAC: 0
Germany
Message 1339485 - Posted: 19 Feb 2013, 23:51:48 UTC - in response to Message 1339480.  

BOINC is using both GPUs ("Config: use all coprocessors" is shown), but only the CUDA part. Don't ask me why it's ignoring the OpenCL parts. ;-)

Gruß,
Gundolf
ID: 1339485 · Report as offensive
Profile arkayn
Volunteer tester
Avatar

Send message
Joined: 14 May 99
Posts: 4438
Credit: 55,006,323
RAC: 0
United States
Message 1339536 - Posted: 20 Feb 2013, 3:33:57 UTC

We have had problems with 7.0.28 recognizing OpenCL.

Try upgrading to the latest beta version which also has a lot of other fixes and see if that fixes the problem.
http://boinc.berkeley.edu/dl/boinc_7.0.52_windows_x86_64.exe

ID: 1339536 · Report as offensive
Profile Mr. Kevvy Crowdfunding Project Donor*Special Project $250 donor
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 3776
Credit: 1,114,826,392
RAC: 3,319
Canada
Message 1339546 - Posted: 20 Feb 2013, 4:11:10 UTC - in response to Message 1339536.  

Try upgrading to the latest beta version which also has a lot of other fixes and see if that fixes the problem


I tried 7.0.52 to fix the problem in my other thread where I don't get AstroPulse 6.04 OpenCL work. Immediately, all my CUDA 6.10 tasks started aborting after a few minutes saying they were taking too long. As soon as I reinstalled 7.0.38 they all started completing normally.

I copied the event log text to paste in here but it was lost. If this isn't a known issue and it would help, I can reinstall it again to capture the abort text.
ID: 1339546 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 1339550 - Posted: 20 Feb 2013, 4:42:06 UTC
Last modified: 20 Feb 2013, 4:46:33 UTC

http://setiathome.berkeley.edu/show_host_detail.php?hostid=6893033
It looks like the machine got 2 AP GPU WUs at 20 Feb 2013, 2:35:24 UTC - 'AstroPulse v6 v6.04 (cuda_opencl_100)'.
Also a few 'SETI@home Enhanced v6.10 (cuda_fermi)' WUs.

http://setiathome.berkeley.edu/apps.php
AFAIK, for pre v7 BOINCs is this (cuda_opencl_100).
For v7 BOINCs is (opencl_nvidia_100), IIRC but also (cuda_opencl_100).

So it looks like the stock v7.0.28 get OpenCL apps/WUs, marked as (cuda_opencl_100).


* Best regards! :-) * Philip J. Fry formerly Sutaru Tsureku, team seti.international founder. * Optimize your PC for higher RAC. * SETI@home needs your help. *
ID: 1339550 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1339562 - Posted: 20 Feb 2013, 5:18:03 UTC - in response to Message 1339546.  
Last modified: 20 Feb 2013, 5:18:59 UTC

Try upgrading to the latest beta version which also has a lot of other fixes and see if that fixes the problem


I tried 7.0.52 to fix the problem in my other thread where I don't get AstroPulse 6.04 OpenCL work. Immediately, all my CUDA 6.10 tasks started aborting after a few minutes saying they were taking too long. As soon as I reinstalled 7.0.38 they all started completing normally.

I copied the event log text to paste in here but it was lost. If this isn't a known issue and it would help, I can reinstall it again to capture the abort text.


See this thread as to why, and how to fix it:

Error

Claggy
ID: 1339562 · Report as offensive

Message boards : Number crunching : NVidea GPU doesn't support openCL message; missing GPU


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.