Skip to content

Instantly share code, notes, and snippets.

@VibhuJawa
Last active October 1, 2019 07:20
Show Gist options
  • Save VibhuJawa/3eaead6caef195fbe2340b71042f954a to your computer and use it in GitHub Desktop.
Save VibhuJawa/3eaead6caef195fbe2340b71042f954a to your computer and use it in GitHub Desktop.
DataPROC nvidia-driver installation
### Problem Diagnosis:
# The current problem is the disparity in the linux-kernel version and
# linux-kernel header version
# we dont have the required linux-header kernel version
# if we upgrade to 4.9.0-11 from 4.9.0-9
# it seems to work
# dont know if that is allowed so stopping work before making more progress
## below scripts allows you to do a successful nvidia-smi installation
### Upgrade linux version
sudo apt-get update --fix-missing
sudo DEBIAN_FRONTEND=noninteractive apt-get -y dist-upgrade
sudo apt-get install linux-headers-4.9.0-11-all -y
sudo reboot
# DIA GPU
apt-get update
apt-get install -y pciutils:
if ! (lspci | grep -q NVIDIA); then
echo 'No NVIDIA card detected. Skipping installation.' >&2
exit 0
fi
# Add non-free Debian 9 Stretch packages.
# See https://www.debian.org/distrib/packages#note
for type in deb deb-src; do
for distro in stretch stretch-backports; do
for component in contrib non-free; do
echo "${type} http://deb.debian.org/debian/ ${distro} ${component}" \
>> /etc/apt/sources.list.d/non-free.list
done
done
done
apt-get update
# Install proprietary NVIDIA Drivers and CUDA
# See https://wiki.debian.org/NvidiaGraphicsDrivers
export DEBIAN_FRONTEND=noninteractive
# apt-get install -y "linux-headers-$(uname -r)"
# Without --no-install-recommends this takes a very long time.
apt-get install -y -t stretch-backports --no-install-recommends \
nvidia-cuda-toolkit nvidia-kernel-common nvidia-driver nvidia-smi
# reate a system wide NVBLAS config
# See http://docs.nvidia.com/cuda/nvblas/
NVBLAS_CONFIG_FILE=/etc/nvidia/nvblas.conf
cat << EOF >> ${NVBLAS_CONFIG_FILE}
# Insert here the CPU BLAS fallback library of your choice.
# The standard libblas.so.3 defaults to OpenBLAS, which does not have the
# requisite CBLAS API.
NVBLAS_CPU_BLAS_LIB /usr/lib/libblas/libblas.so
# Use all GPUs
NVBLAS_GPU_LIST ALL
# Add more configuration here.
EOF
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment