Hi
I’ve just installed the cuda toolkit and the sdk and sucefully compiled the sdk examples on a Debian 5.0.2 with a Geforce 8600 GT. But when runing, I get segmentations faults. Here’s an example
me@pc:~/NVIDIA_GPU_Computing_SDK/C/bin/linux/release$ ./particles
[CUDA Particles Simulation]
grid: 64 x 64 x 64 = 262144 cells
particles: 32768
Segmentation fault
I have no clue on where to start.
Thank you
I’ve exactly the same problem, all the SDK examples are returning segmentation fault.
I have no clue, but I install SDK and nvcc as user on my home…
Thanks !
I had the same problem. For me it was because the installed driver and the CUDA version were slightly incompatible. Maybe installing the developer driver and the corresponding toolkit+sdk from the nVidia site might help.
I don’t know why, but I think it must have some relations with the NVIDIA driver version. I have tried my best to recompile the CUDA 2.3 in my GPU cluster, but it doesn’t work for me, and I check the results of deviceQuery and find that it cannot find CUDA device. So, I get the same results with you all.