I have been trying to get cuda installed using the instruction guide. but I always get error message from nvidia-smi. The following are the sequence of command I tried. Can anyone explain why I can’t get it to work? Thanks,
1 gcc
2 sudo apt install gcc
3 lspci|grep -i nvidia
4 uname -m && cat /etc/*release
5 gcc --version
6 uname -r
7 sudo apt-get install linux-headers-$(uname -r)
10 sudo dpkg -i cuda-repo-ubuntu1804_10.1.105-1_amd64.deb
11 sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
12 sudo apt-get update
13 sudo apt-get install cuda-10-1
14 sudo nano /etc/modprobe.d/blacklist-nouveau.conf
15 sudo update-initramfs -u
16 reboot
then I changed $PATH and $LD_LIBRARY_PATH and verified nvcc version. But nvidia-smi is giving me the following errors.
NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.
later I re installed nvidia driver, but still got the same error. anyone knows why?
71 dpkg -l |grep nvidia
72 sudo dpkg -purge nvidia-*
73 sudo apt-get --purge remove cuda
74 sudo apt-get --purge remove nvidia*
75 nvidia-smi
76 reboot
77 sudo add-apt-repository ppa:graphics-drivers/ppa
78 sudo apt-get update
79 sudo apt-get install nvidia-driver-418
81 reboot
82 nvidia-smi
NVIDIA-SMI has failed because it couldn’t communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.