I have two GTX 480s. When I run GPUGRID only Device 0 is active. GPU Grid will not start on the other Video Card. If I run SETI, both Video cards are used so I am sure I have BOINC Configured correctly. Boinc Version 7.0.64

5/11/2013 1:46:10 PM | | CUDA: NVIDIA GPU 0: GeForce GTX 480 (driver version 314.22, CUDA version 5.0, compute capability 2.0, 1536MB, 1424MB available, 1345 GFLOPS peak)
5/11/2013 1:46:10 PM | | CUDA: NVIDIA GPU 1: GeForce GTX 480 (driver version 314.22, CUDA version 5.0, compute capability 2.0, 1536MB, 1424MB available, 1345 GFLOPS peak)
5/11/2013 1:46:10 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX 480 (driver version 314.22, device version OpenCL 1.1 CUDA, 1536MB, 1424MB available, 1345 GFLOPS peak)
5/11/2013 1:46:10 PM | | OpenCL: NVIDIA GPU 1: GeForce GTX 480 (driver version 314.22, device version OpenCL 1.1 CUDA, 1536MB, 1424MB available, 1345 GFLOPS peak)

Use All gpu option active in cc config?

---

DSKAG Austria Research Team: http://www.research.dskag.at

Click on the link below and see if this helps:



http://www.gpugrid.net/forum_thread.php?id=3136&nowrap=true#26948



The "<use_all_gpus> 1 </ use_all_gpus>" must be in your cc_config.xml file.

This is the cc_config.xml I am using


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

It works for SETI. When I run SETI it runs on both Video cards with no problem. For some reason, GPUGRID will only work on one Video Card.

Is that your full cc_config file or do you have an exclude GPU somewhere?
Do you have any other bespoke settings; app_config or app_info?

Perhaps you could post the first page of your Boinc Manager Event Log, try a project reset and a system restart. If that fails I would suggest you do a clean driver install and then reinstall Boinc or move from 7.0.28 to 7.0.64.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Is that your full cc_config file or do you have an exclude GPU somewhere?
Do you have any other bespoke settings; app_config or app_info?

Perhaps you could post the first page of your Boinc Manager Event Log, try a project reset and a system restart. If that fails I would suggest you do a clean driver install and then reinstall Boinc or move from 7.0.28 to 7.0.64.

I have upgraded to 7.0.64 and am running latest NVIDIA Drivers. I do not think I have an app_config or app_info. Here is my log.

5/13/2013 4:35:37 PM | | Starting BOINC client version 7.0.64 for windows_x86_64
5/13/2013 4:35:37 PM | | log flags: file_xfer, sched_ops, task
5/13/2013 4:35:37 PM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
5/13/2013 4:35:37 PM | | Data directory: E:\ProgramData\BOINC
5/13/2013 4:35:37 PM | | Running under account Mitch
5/13/2013 4:35:37 PM | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7 CPU 975 @ 3.33GHz [Family 6 Model 26 Stepping 5]
5/13/2013 4:35:37 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 popcnt syscall nx lm vmx tm2 pbe
5/13/2013 4:35:37 PM | | OS: Microsoft Windows 7: Home Premium x64 Edition, Service Pack 1, (06.01.7601.00)
5/13/2013 4:35:37 PM | | Memory: 5.99 GB physical, 11.98 GB virtual
5/13/2013 4:35:37 PM | | Disk: 16.25 GB total, 6.22 GB free
5/13/2013 4:35:37 PM | | Local time is UTC -5 hours
5/13/2013 4:35:37 PM | | CUDA: NVIDIA GPU 0: GeForce GTX 480 (driver version 314.22, CUDA version 5.0, compute capability 2.0, 1536MB, 1308MB available, 1345 GFLOPS peak)
5/13/2013 4:35:37 PM | | CUDA: NVIDIA GPU 1: GeForce GTX 480 (driver version 314.22, CUDA version 5.0, compute capability 2.0, 1536MB, 1423MB available, 1345 GFLOPS peak)
5/13/2013 4:35:37 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX 480 (driver version 314.22, device version OpenCL 1.1 CUDA, 1536MB, 1308MB available, 1345 GFLOPS peak)
5/13/2013 4:35:37 PM | | OpenCL: NVIDIA GPU 1: GeForce GTX 480 (driver version 314.22, device version OpenCL 1.1 CUDA, 1536MB, 1423MB available, 1345 GFLOPS peak)
5/13/2013 4:35:37 PM | | Config: use all coprocessors
5/13/2013 4:35:37 PM | Collatz Conjecture | URL http://boinc.thesonntags.com/collatz/; Computer ID 27359; resource share 100
5/13/2013 4:35:37 PM | Milkyway@Home | URL http://milkyway.cs.rpi.edu/milkyway/; Computer ID 126422; resource share 100
5/13/2013 4:35:37 PM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 5442522; resource share 100
5/13/2013 4:35:37 PM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 54421; resource share 100
5/13/2013 4:35:37 PM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 1082205; resource share 100
5/13/2013 4:35:37 PM | World Community Grid | General prefs: from World Community Grid (last modified 04-May-2013 11:16:01)
5/13/2013 4:35:37 PM | World Community Grid | Computer location: work
5/13/2013 4:35:37 PM | | General prefs: using separate prefs for work
5/13/2013 4:35:37 PM | | Reading preferences override file
5/13/2013 4:35:37 PM | | Preferences:
5/13/2013 4:35:37 PM | | max memory usage when active: 6135.14MB
5/13/2013 4:35:37 PM | | max memory usage when idle: 6135.14MB
5/13/2013 4:35:37 PM | | max disk usage: 6.79GB
5/13/2013 4:35:37 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)
5/13/2013 4:35:37 PM | | Not using a proxy

I'm having the same issue... Was this ever fixed?

There are some fixes in the development versions, but not in the 7.0.64 version.

I've experienced numerous instances where one or both of my GPU's stops running WU's or Boinc doesn't fetch work. There seems to be several problems. It's really annoying to have two GPU's stop crunching just so a few CPU WU's can crunch. I've also had the odd situation where I've had some cached CPU work units and Boinc wouldn't fetch work for the GPUs (for ages). The quick-fix, reduce the cache and ask for work. Doesn't really make much sense to have 3days of cached CPU tasks, and need to reduce the cache to 0.25 to get GPU work.

Anyway, if you keep a low cache, a small project list (or suspend some projects) and keep GPUGrid's project Resource Share very high it should help.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help