I can “make” the SDK sample files with no errors. But when I change into “…linux/release” directory and try to execute them I get the message “…cannot execute the binary file”. It seems as though nvcc is not activated. Also when I type in “nvidia-settings” to see how the X-Driver is configured, I get the message “you do not appear to be using the X-Driver”. I have exited the gui with ctrl-alt-bkspac, typed /sbin/init 3 and then startx which brings me back to the GUI and there is no difference. I have the “190.18” cudadrivers and toolkits installed (version 2.3). I have tried the older versions of the drivers (2.2,2.1). I am running RedHat Linux Enterprise version 5.3. I have 4 C1060 cards. The motherboard is Tyan S7025. I have 24GB of RAM. I have never been able to get nvcc -v to tell me what version I am running. There must be a simple fix, but I can’t find it.
Well - probably you should check before installing the cudadriver190.18 that older ones are completly uninstalled. Then reinstall the cudadriver.
Have you a proper enviroment setting? This could be achieved by adding the following to your .bashrc file in your /home:
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
The /usr/local/cuda/lib64 is appropriate for 64bit Systems. For 32 bit: /usr/local/cuda/lib
Then you open a new bash or you do a:
source ~/bashrc