CUDA 11.1 Installation on Ubuntu 18.04

cuda_11.1_installation_on_Ubuntu_18.04.md

This gist contains instructions about cuda v11.1 and cudnn 8.0.4 installation in Ubuntu 18.04 for Tensorflow 2.4.1

• If you have previous installation remove it first.

!/bin/bash
sudo apt-get purge *nvidia*
sudo apt remove --autoremove nvidia-*
sudo rm /etc/apt/sources.list.d/cuda*
sudo apt remove --autoremove nvidia-cuda-toolkit
sudo apt-get autoremove && sudo apt-get autoclean
sudo rm -rf /usr/local/cuda*

• Verify your gpu cuda

lspci | grep -i nvidia

• gcc compiler is required for development using the cuda toolkit. to verify the version of gcc install enter

gcc --version

• System update

sudo apt-get update
sudo apt-get upgrade

• Install other import packages

sudo apt-get install g++ freeglut3-dev build-essential libx11-dev libxmu-dev libxi-dev libglu1-mesa libglu1-mesa-dev

• First get the PPA repository driver

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
echo "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64 /" | sudo tee /etc/apt/sources.list.d/cuda.list
sudo apt-get update

• Installing CUDA-11.1

sudo apt-get -o Dpkg::Options::="--force-overwrite" install cuda-11-1 cuda-drivers

• Setup your paths

echo 'export PATH=/usr/local/cuda-11.1/bin:$PATH' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH=/usr/local/cuda-11.1/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
source ~/.bashrc
sudo ldconfig

• Install cuDNN v8.0.4

CUDNN_TAR_FILE="cudnn-11.1-linux-x64-v8.0.4.30.tgz"
wget https://developer.nvidia.com/compute/machine-learning/cudnn/secure/8.0.4/11.1_20200923/cudnn-11.1-linux-x64-v8.0.4.30.tgz
tar -xzvf ${CUDNN_TAR_FILE}

• Copy the following files into the cuda toolkit directory.

sudo cp -P cuda/include/cudnn*.h /usr/local/cuda-11.1/include
sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda-11.1/lib64/
sudo chmod a+r /usr/local/cuda-11.1/lib64/libcudnn*

• Finally, to verify the installation, check (maybe the Cuda version in nvidia-smi and nvcc is different)

sudo apt install nvidia-cuda-toolkit
nvidia-smi
nvcc -V

• Install Tensorflow 2.4.1 (tf-nightly-gpu if needed) because it is stable and compatible with CUDA 11.1 Toolkit and cuDNN 8.0.4

pip3 install --user tf-nightly-gpu
pip3 install --user tensorflow==2.4.1
conda install -c anaconda cudnn
conda install -c anaconda cudatoolkit

11.1 not CUDA 11.0

OS: Ubuntu 18.04 LTS (VM-based installation)
Machine: HP Victus (with Nvidia RTX 3060)

wget http://people.cs.uchicago.edu/~kauffman/nvidia/cudnn/cudnn-11.1-linux-x64-v8.0.4.30.tgz

throws this error

Connecting to people.cs.uchicago.edu (people.cs.uchicago.edu)|128.135.164.80|:80... connected.
HTTP request sent, awaiting response... 403 Forbidden
2022-03-28 04:34:38 ERROR 403: Forbidden.

Please help me solve this.

OS: Ubuntu 18.04 LTS (VM-based installation) Machine: HP Victus (with Nvidia RTX 3060)

wget http://people.cs.uchicago.edu/~kauffman/nvidia/cudnn/cudnn-11.1-linux-x64-v8.0.4.30.tgz

throws this error

Connecting to people.cs.uchicago.edu (people.cs.uchicago.edu)|128.135.164.80|:80... connected.
HTTP request sent, awaiting response... 403 Forbidden
2022-03-28 04:34:38 ERROR 403: Forbidden.

Please help me solve this.

Oh, that direct link for cuDNN 8.0.4 is removed. You can download it directly from Nvidia at: https://developer.nvidia.com/rdp/cudnn-archive
I also fixed the direct link in the gist.
Thank you.

A small correction :
sudo cp -P cuda/include/cudnn.h /usr/local/cuda-11.1/include
should be
sudo cp -P cuda/include/cudnn*.h /usr/local/cuda-11.1/include
to copy all the headers.

A small correction : sudo cp -P cuda/include/cudnn.h /usr/local/cuda-11.1/include should be sudo cp -P cuda/include/cudnn*.h /usr/local/cuda-11.1/include to copy all the headers.

Thank you very much for your correction!