InstallingCUDA 8.0 on a fresh installation of Ubuntu 16.0
(tested on Alienware-17-R4，i7 7820hk，Gtx-1080)
1. Launch [Software& Updates]. Select [Additional Drivers] tab.
2. In the list, findyour graphic card. Among the drivers that can be used for the card, choose theproprietary driver from NVIDIA. Then press [Apply Changes] button. In my case,under the graphics card name "NVIDIA Corporation: GM107M [Geforce GTX950M]", there were two selections:
· Using NVIDIAbinary driver - version 375.66 from nvidia-375 (proprietary, tested)
· Using X.Org Xserver - Nouveau display driver from xserver-xorg-video-nouveau (open source)
3. Delete defaultinstalled video drivers with $ sudoapt remove xserver-xorg-video*.
5. Download CUDA8.0 Toolkit from here. Among the installer types, choose "runfile (local)". Thisdownloads cuda_8.0.61_375.26_linux.run.
6. Run theinstaller with $ sudo shcuda_8.0.61_375.26_linux.run.
· [Optional] Ifyour currently installed NVIDIA driver version is higher than the driverversion contained in the downloaded installer, you can choose not to installthe driver while installing CUDA. In my case, since I already have driverversion 375.66, which is higher than 375.26contained in the installer, I chose not to install.
7. After install,config your binary path and library path (You can follow the directions fromthe instller). If you choose to configure ld.so.conf and the following erroroccurs: libEGL.so.1 is not asymbolic link, follow the direction from this link.
8.The laststep of installation CUDA is to add the path of it to source bash file:
The the twofollowing line must be added to the file with regard to CUDA 8.0 installation:
Check Instructions：env #check environment variable
nvcc–V #check cuda tools release version