Ctrl+K

Install

Contents

Install#

JAX-AM supports Linux and macOS. Create a new conda environment and run

git clone https://github.com/tianjuxue/jax-am.git
cd jax-am
pip install .

Note: JAX-AM depends on petsc4py. We have found difficulty installing petsc4py with pip on certain platforms. Installing petsc4py with conda is therefore recommended (see here).

Several remarks:

Conda#

Conda is an open source package management system and environment management system.The installer is offered as a shell script on here. Find and download the version that matches your system and CPU architecture. Then make the script you download executable.

wget https://repo.anaconda.com/archive/Anaconda3-2023.03-1-Linux-x86_64.sh
chmod -v +x Anaconda3-2022.05-Linux-x86_64.sh

Once done, execute the script to run the installer for Anaconda.

./Anaconda3-2023.03-1-Linux-x86_64.sh

Then you will see that the installation program has an EULA, and you need to press “Enter” to agree. This will install Anaconda in the default location, which is ./~/anaconda3. Finally, to complete the installation of Anaconda, update the PATH variable. Open your .bashrc file using a text editor (e.g. Vim, Nano, VSCode, etc.), add the following three lines, and save the file.

if ! [[ $PATH =~ "$HOME/anaconda3/bin" ]]; then
    PATH="$HOME/anaconda3/bin:$PATH"
fi

Restart the terminal and run the following command to check whether Anaconda has been successfully installed and set in PATH.

conda list

Anaconda is successfully installed if you get a list of packages. To create and open an environment, run

conda create -n myenv python=3.9
conda activate myenv

CUDA#

Compute Unified Device Architecture (CUDA) is a parallel computing platform and application programming interface developed by Nvidia, which enables software developers to perform general-purpose computing using GPUs that support CUDA software. Before using JAX, you need to install CUDA first.

In general, the process of installing and using CUDA under Linux is as follows:

  1. Preparation for installation of CUDA Toolkit

  2. Install the CUDA Toolkit.

To verify that your GPU is CUDA-capable, run

lspci | grep -i nvidia

You will get a prompt similar to the following if your GPU supports CUDA programming.

3b:00.0 VGA compatible controller: NVIDIA Corporation Device 2230 (rev a1)
3b:00.1 Audio device: NVIDIA Corporation Device 1aef (rev a1)

Make sure that gcc, g++ and make are installed on your system. If you want to run CUDA examples, you need to install the appropriate dependency libraries.

sudo apt update # update apt
sudo apt install gcc g++ make # install gcc, g++, make
sudo apt install libglu1-mesa libxi-dev libxmu-dev libglu1-mesa-dev freeglut3-dev # install dependency libraries

Select the installation method from the CUDA toolbox that matches your system version and architecture download page, download and run the runfile.

# Download CUDA runfile
wget https://developer.download.nvidia.com/compute/cuda/12.1.1/local_installers/cuda_12.1.1_530.30.02_linux.run 
# Install CUDA Toolkit
sudo sh cuda_12.1.1_530.30.02_linux.run

Note: You can get information about your system’s architecture, version, etc. by run

uname -m && cat /etc/*release

After the CUDA toolkit has been installed, the screen will output the following message:

Driver:   Installed
Toolkit:  Installed in /usr/local/cuda-10.1/
Samples:  Installed in /home/abneryepku/

Please make sure that
 -   PATH includes /usr/local/cuda-10.1/
 -   LD_LIBRARY_PATH includes /usr/local/cuda-10.1/lib64, or, add /usr/local/cuda-10.1/lib64 to /etc/ld.so.conf and run ldconfig as root

Write the following text in the ~/.bashrc file to complete the CUDA configuration.

export PATH=$PATH:/usr/local/cuda-10.1/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-10.1/lib64:/usr/lib/x86_64-linux-gnu

You can check if CUDA is successfully installed by run

nvcc --version

If the installation is successful, you will receive a message similar to the following

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2023 NVIDIA Corporation
Built on Tue_Feb__7_19:32:13_PST_2023
Cuda compilation tools, release 12.1, V12.1.66
Build cuda_12.1.r12.1/compiler.32415258_0

cuDNN#

The NVIDIA CUDA Deep Neural Network library (cuDNN) is a GPU-accelerated library of primitives for deep neural networks. You need to install it after installing CUDA. Select the cuDNN version that matches your CUDA version from the download page.

The installation process is actually copying the cuDNN header files into the CUDA header directory. Unzip the downloaded cuDNN, and then execute the following command

# Copy cuDNN files
sudo cp cuda/include/* /usr/local/cuda-12.0/include/
sudo cp cuda/lib64/* /usr/local/cuda-12.0/lib64/
# Add executable permissions
sudo chmod +x /usr/local/cuda-12.0/include/cudnn.h
sudo chmod +x /usr/local/cuda-12.0/lib64/libcudnn*

You can run the following command to check if cuDNN has been installed successfully.

cat /usr/local/cuda-12.0/include/cudnn.h | grep CUDNN_MAJOR -A 2

Neper#

Neper is a free / open source software package for polycrystal generation and meshing. It can be used to generate polycrystals with a wide variety of morphological properties. Install the following components before downloading Neper.

sudo apt install cmake
sudo apt-get install libnlopt.dev
sudo apt-get install povray
sudo apt-get install libgmsh-dev
sudo apt-get install imagemagick
sudo apt-get install libgsl0-dev

The latest version is available on the main branch of the GitHub repository. To get the latest version as a Git user, run:

git clone https://github.com/neperfepx/neper.git

Enter the neper/src/ directory, create a folder named build, and execute the compilation.

mkdir build && cd buid && cmake ..
make -j
sudo make install

You can perform testing by running the following commands to check if Neper has been successfully installed.

neper -T -n 10 -o test
neper -V  test.tess -datacellcol id -print img_test

If Neper has been successfully installed, you should be able to find two files named test.tess and img_test.png in the current directory where you executed the above commands.

Contents