SDK Examples Hang - CUDA 3.2
Hi all,

I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)

When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.

Any ideas?

Thanks,
Stefano
Hi all,



I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)



When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.



Any ideas?



Thanks,

Stefano

#1
Posted 11/12/2010 02:01 AM   
Hi all,

I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)

When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.

Any ideas?

Thanks,
Stefano
Hi all,



I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)



When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.



Any ideas?



Thanks,

Stefano

#2
Posted 11/12/2010 02:01 AM   
No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.

deviceQuery used to work when there were just 2 C2050s, making this even stranger...

64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK

[quote name='zeus13i' date='12 November 2010 - 02:01 AM' timestamp='1289527315' post='1145282']
Hi all,

I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)

When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.

Any ideas?

Thanks,
Stefano
[/quote]
No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.



deviceQuery used to work when there were just 2 C2050s, making this even stranger...



64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK



[quote name='zeus13i' date='12 November 2010 - 02:01 AM' timestamp='1289527315' post='1145282']

Hi all,



I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)



When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.



Any ideas?



Thanks,

Stefano

#3
Posted 11/12/2010 03:18 PM   
No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.

deviceQuery used to work when there were just 2 C2050s, making this even stranger...

64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK

[quote name='zeus13i' date='12 November 2010 - 02:01 AM' timestamp='1289527315' post='1145282']
Hi all,

I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)

When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.

Any ideas?

Thanks,
Stefano
[/quote]
No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.



deviceQuery used to work when there were just 2 C2050s, making this even stranger...



64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK



[quote name='zeus13i' date='12 November 2010 - 02:01 AM' timestamp='1289527315' post='1145282']

Hi all,



I have a fresh install of 64-bit Ubuntu 10.04 with CUDA 3.2 (cudatoolkit_3.2.12_linux_64_ubuntu10.04.run), 260.19.14 drivers (devdriver_3.2_linux_64_260.19.14.run) and 3.2 SDK (gpucomputingsdk_3.2.12_linux.run). All SDK examples compiled successfully. 1 GPU currently attached: C2050 (which nvidia-settings picks up fine)



When I run deviceQuery, amongst other programs, the app(s) seem to just hang. No output, no errors. Interestingly, if I run deviceQueryDrv then it works as expected. Note, too, that I'm able to run some of the OpenGL examples, such as nbody.



Any ideas?



Thanks,

Stefano

#4
Posted 11/12/2010 03:18 PM   
[quote name='slayton' date='12 November 2010 - 03:18 PM' timestamp='1289575107' post='1145540']
No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.

deviceQuery used to work when there were just 2 C2050s, making this even stranger...

64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK
[/quote]

64-bit Ubuntu 12.04 beta 2 with kernel 3.2.0-23-generic
dmesg: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 285.05.33

installed from the following files:
[list]
[*]cudatoolkit_4.1.28_linux_64_ubuntu11.04.run
[*]gpucomputingsdk_4.1.28_linux.run
[*]NVIDIA-Linux-x86_64-285.05.33.run
[/list]

I realize that the 4.1.28 toolkit package mentioned ubuntu version 11.04 and not 12.04, but since you experienced similar problems it may not be that.

deviceQuery now works fine by the way, but [b]nbody -benchmark[/b] and everything else I tried just hangs.

Michael
[quote name='slayton' date='12 November 2010 - 03:18 PM' timestamp='1289575107' post='1145540']

No solutions, but another datapoint here -- GTX 295 & 2 Tesla C2050s, deviceQuery gives no output, I don't actually have a devicequeryDrv (no errors building though). Can also run (at least) nbody, fluidsGL and simpleMPI.



deviceQuery used to work when there were just 2 C2050s, making this even stranger...



64-bit Ubuntu 10.04, 260.29.21 drivers, all 4 devices show up in nvidia-settings (in a strange order -- C2050, GTX295(1), GTX295(2), C2050), 3.2 SDK





64-bit Ubuntu 12.04 beta 2 with kernel 3.2.0-23-generic

dmesg: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 285.05.33



installed from the following files:


  • cudatoolkit_4.1.28_linux_64_ubuntu11.04.run
  • gpucomputingsdk_4.1.28_linux.run
  • NVIDIA-Linux-x86_64-285.05.33.run




I realize that the 4.1.28 toolkit package mentioned ubuntu version 11.04 and not 12.04, but since you experienced similar problems it may not be that.



deviceQuery now works fine by the way, but nbody -benchmark and everything else I tried just hangs.



Michael

#5
Posted 04/18/2012 07:42 PM   
Scroll To Top