no CUDA-capable device is detected

I ran ./deviceQuery program and it failed. I’m not sure what I’m missing here. Could it be the GPUs installed are not CUDA capable?

This is the output of the the deviceQuery:
./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 38
-> no CUDA-capable device is detected
Result = FAIL

I ran the deviceQuery on Ubuntu 14.04.4 LTS. The GPUs installed are:
lspci | grep -i nvidia
06:00.0 VGA compatible controller: NVIDIA Corporation GK104GL [GRID K2] (rev a1)
07:00.0 VGA compatible controller: NVIDIA Corporation GK104GL [GRID K2] (rev a1)

The driver is:
cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 352.63 Sat Nov 7 21:25:42 PST 2015
GCC version: gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)

Thanks in advance for any help.

Have you installed nvidia-modprobe?
I had to add the following line to /etc/rc.d./rc.local to get CUDA devices to work:

# Create missing nvidia device nodes after reboot
/usr/bin/nvidia-modprobe -c 0 -u

Mats

I don’t have root access. I checked for /usr/bin/nvidia-modprobe and it is not there. I guess it is not installed.

Okay,
I have:

bash-4.3# ls -l /usr/bin/ |grep nvidia
-rwxr-xr-x   1 root   root      23K feb 12 02:27 nvidia-bug-report.sh*
-rwxr-xr-x   1 root   root     204K feb 12 02:27 nvidia-debugdump*
-rwxr-xr-x   1 root   root     300K feb 12 02:27 nvidia-installer*
-rwxr-xr-x   1 root   root      26K feb 12 02:28 nvidia-modprobe*
-rwxr-xr-x   1 root   root      33K feb 12 02:29 nvidia-persistenced*
-rwxr-xr-x   1 root   root     1,8M feb 12 02:28 nvidia-settings*
-rwxr-xr-x   1 root   root     245K feb 12 02:27 nvidia-smi*
lrwxrwxrwx   1 root   root       16 nov 16 19:12 nvidia-uninstall -> nvidia-installer*
-rwxr-xr-x   1 root   root     170K feb 12 02:28 nvidia-xconfig*

Do you see these three devices?:

bash-4.3# ls -l /dev |grep nvidia
crw-rw-rw-  1 root root    245,   0 mar  9 01:28 nvidia-uvm
crw-rw-rw-  1 root root    195,   0 mar  9 01:29 nvidia0
crw-rw-rw-  1 root root    195, 255 mar  9 01:29 nvidiactl

No, I don’t have the three devices that you listed.
This is what I have:

ls -l /usr/bin/ |grep nvidia
lrwxrwxrwx  1 root   root          52 Mar  7 10:37 nvidia-bug-report.sh -> /etc/alternatives/x86_64-linux-gnu_nvidia_bug_report*
lrwxrwxrwx  1 root   root          58 Mar  7 10:37 nvidia-cuda-mps-control -> /etc/alternatives/x86_64-linux-gnu_nvidia-cuda-mps-control*
lrwxrwxrwx  1 root   root          57 Mar  7 10:37 nvidia-cuda-mps-server -> /etc/alternatives/x86_64-linux-gnu_nvidia-cuda-mps-server*
lrwxrwxrwx  1 root   root          51 Mar  7 10:37 nvidia-debugdump -> /etc/alternatives/x86_64-linux-gnu_nvidia-debugdump*
-rwxr-xr-x  1 root   root         270 Aug  4  2015 nvidia-detector*
lrwxrwxrwx  1 root   root          54 Mar  7 10:37 nvidia-persistenced -> /etc/alternatives/x86_64-linux-gnu_nvidia_persistenced*
-rwxr-xr-x  1 root   root      204064 Apr 15  2015 nvidia-settings*
lrwxrwxrwx  1 root   root          45 Mar  7 10:37 nvidia-smi -> /etc/alternatives/x86_64-linux-gnu_nvidia_smi*
lrwxrwxrwx  1 root   root          49 Mar  7 10:37 nvidia-xconfig -> /etc/alternatives/x86_64-linux-gnu_nvidia_xconfig*
-rwxr-xr-x  1 root   root          99 Nov 16 13:50 start-nvidia-persistenced*
-rwxr-xr-x  1 root   root         109 Nov 16 13:50 stop-nvidia-persistenced*

The administrator installed another version of the driver and it is working now. I can see the nvidia-modprobe. I guess the problem was with the driver version.

The current driver:

cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module  352.79  Wed Jan 13 16:17:53 PST 2016
GCC version:  gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)

The previous driver:

cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module 352.63 Sat Nov 7 21:25:42 PST 2015
GCC version: gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)

Thanks
Rami

Hi, I am having the same issue with deviceQuery under Cuda 6.5, although I have the 352.79 driver.

When ran with root privileges, deviceQuery Fails:

./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 38
-> no CUDA-capable device is detected
Result = FAIL

My NVIDIA driver:

NVRM version: NVIDIA UNIX x86_64 Kernel Module 352.79 Wed Jan 13 16:17:53 PST 2016
GCC version: gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.1)

Which version of Cuda works best with this driver? I was last able to run code that used Cuda successfully in early March 2016.

Thank you!

Octavia