Hi Team

Upgraded to the latest BOINC, restarted, and its decided to use just the second of my two GPU's, where as before it was using both without manual intervention.

Here is part of the Event log:

Sun 16 Feb 2014 23:16:43 CST | | Starting BOINC client version 7.2.39 for x86_64-pc-linux-gnu
Sun 16 Feb 2014 23:16:43 CST | | log flags: file_xfer, sched_ops, task
Sun 16 Feb 2014 23:16:43 CST | | Libraries: libcurl/7.32.0 OpenSSL/1.0.1e zlib/1.2.8 libidn/1.28 librtmp/2.3
Sun 16 Feb 2014 23:16:43 CST | | Data directory: /home/mark/Apps/BOINC
Sun 16 Feb 2014 23:16:43 CST | | CUDA: NVIDIA GPU 0 (not used): GeForce GTX 580 (driver version unknown, CUDA version 5.5, compute capability 2.0, 1535MB, 654MB available, 1628 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | CUDA: NVIDIA GPU 1: GeForce GTX 570 (driver version unknown, CUDA version 5.5, compute capability 2.0, 1280MB, 1218MB available, 1498 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | OpenCL: NVIDIA GPU 0 (ignored by config): GeForce GTX 580 (driver version 319.32, device version OpenCL 1.1 CUDA, 1535MB, 654MB available, 1628 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | OpenCL: NVIDIA GPU 1: GeForce GTX 570 (driver version 319.32, device version OpenCL 1.1 CUDA, 1280MB, 1218MB available, 1498 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | Host name: zen
Sun 16 Feb 2014 23:16:43 CST | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
Sun 16 Feb 2014 23:16:43 CST | | 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 ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 popcnt tsc_deadline_timer aes xsave avx lahf_lm ida arat epb xsaveopt pln pts dtherm tpr_shadow vnmi flexpriority ept vpid
Sun 16 Feb 2014 23:16:43 CST | | OS: Linux: 3.11.0-15-generic
Sun 16 Feb 2014 23:16:43 CST | | Memory: 15.64 GB physical, 17.32 GB virtual
Sun 16 Feb 2014 23:16:43 CST | | Disk: 2.69 TB total, 848.12 GB free
Sun 16 Feb 2014 23:16:43 CST | | Local time is UTC +10 hours

Soooo any idea how to fix as well as how to change Boinc to follow its original behaviour in future versions?

Cheers

Mark

Sun 16 Feb 2014 23:16:43 CST | | Starting BOINC client version 7.2.39 for x86_64-pc-linux-gnu
Sun 16 Feb 2014 23:16:43 CST | | log flags: file_xfer, sched_ops, task
Sun 16 Feb 2014 23:16:43 CST | | Libraries: libcurl/7.32.0 OpenSSL/1.0.1e zlib/1.2.8 libidn/1.28 librtmp/2.3
Sun 16 Feb 2014 23:16:43 CST | | Data directory: /home/mark/Apps/BOINC
Sun 16 Feb 2014 23:16:43 CST | | CUDA: NVIDIA GPU 0 (not used): GeForce GTX 580 (driver version unknown, CUDA version 5.5, compute capability 2.0, 1535MB, 654MB available, 1628 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | CUDA: NVIDIA GPU 1: GeForce GTX 570 (driver version unknown, CUDA version 5.5, compute capability 2.0, 1280MB, 1218MB available, 1498 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | OpenCL: NVIDIA GPU 0 (ignored by config): GeForce GTX 580 (driver version 319.32, device version OpenCL 1.1 CUDA, 1535MB, 654MB available, 1628 GFLOPS peak)
Sun 16 Feb 2014 23:16:43 CST | | OpenCL: NVIDIA GPU 1: GeForce GTX 570 (driver version 319.32, device version OpenCL 1.1 CUDA, 1280MB, 1218MB available, 1498 GFLOPS peak)

Soooo any idea how to fix as well as how to change Boinc to follow its original behaviour in future versions?

By default Boinc will only use the most capable GPU, it hasn't changed with Boinc 7.2.39,
perhaps your GPUs available memory has changed, or one of their speeds is slightly different now, (I think they have to be within 10 or 20% to be similar enough)

http://boinc.berkeley.edu/dev/forum_thread.php?id=7940&postid=46063

By default BOINC uses only the best GPU. Even two gpus of the same make and model, can still differ. BOINC decides which gpu is best based on these factors, in decreasing priority):
- compute capability
- software version
- available memory
- speed

To use all GPUs put the following in your cc_config.xml:

<use_all_gpus>1</use_all_gpus>

http://boinc.berkeley.edu/wiki/Client_configuration

Claggy

Hi Claggy

Thanks for that, it did the trick .... but now I am wondering how they were both working that way in the past... *scratches head* lol

I should add the main GPU is an nVidia GTX 580 (The main Video Card) and the second is an nVidia GTX 570 (Not driving a display). It may have just come down to the amount of usable ram on them both as the main one is running the display - though that is only a 300MB load, maybe its just enough now for it to decide that the second one is to be used and to ignore the first. Anyway all fixed now. Thanks again.

Cheers

Mark

---