Error compiling the SDK examples

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello,

When compiling the SDK examples I get this error:

NVIDIA: could not open the device file /dev/nvidiactl (No such device or address).
Using device 0: Device Emulation (CPU)
cufft: ERROR: /root/cuda-stuff/sw/rel/gpgpu/toolkit/r2.1/cufft/src/config.cu, line 138
cufft: ERROR: CUFFT_INTERNAL_ERROR
cufft: ERROR: /root/cuda-stuff/sw/rel/gpgpu/toolkit/r2.1/cufft/src/config.cu, line 138
cufft: ERROR: CUFFT_INTERNAL_ERROR
Xlib: extension “GLX” missing on display “:0.0”.
freeglut (/home/NVIDIA_CUDA_SDK/bin/linux/release/fluidsGL): OpenGL GLX extension not supported by display ‘:0.0’

Can anyone help me please?

Thanks a lot!

(I have a 9600GT graphiocs card and I run a 8.04 Ubuntu system. I installed the 173.14.12 nvidia driver using Envy)

2

Thank-you anyway, but it seems that after running ‘nvidia-xconfig’ and restarting the x server, they seem to work. However, when running ‘make’ in the SDK directory, I get this error

Makefile:34: *** commands commence before first target. Stop.
make[1]: Leaving directory `/home/NVIDIA_CUDA_SDK/projects/eclCtst2’
make: *** [projects/eclCtst2/Makefile.ph_clean] Error 2

Does anyone have any ideea why?

3

You might have to map your library and bin paths to your cuda install folder. I will post the code for it when I get back to work. But you can find an example of how to do it in the CUDA release document.

4

Here you go it is like this

PATH=$PATH:/LocationOfCudaLibraries/bin

LD_LIBRARY=$LD_LIBRARY:/LocationOfCudaLibraries/lib

export PATH

export LD_Library