Installation Guide

Installation Guide#

Introduction#

cuTAGI is a high-performance C++/CUDA implementation of the TAGI (Tractable Approximate Gaussian Inference) method. Python users can utilize pytagi, a Python wrapper around the cuTAGI backend, for seamless integration within Python environments.

This guide covers detailed instructions on installing both pytagi and cuTAGI on various platforms.

Prerequisites for Local Installation#

  • Compiler: C++14 support

  • CMake: Version >= 3.23

  • CUDA Toolkit: Optional for GPU support

  • Python Version: Python >= 3.9 (Python 3.10 recommended)

Installing pytagi#

pytagi is the Python interface for the cuTAGI backend.

Setting Up a Conda Environment#

  1. Install Miniconda: Follow the official instructions.

  2. Create a New Conda Environment:

    conda create --name your_env_name python=3.10

  3. Activate the Environment:

    conda activate your_env_name

PyPI Installation#

  1. Ensure Conda Environment is Activated.

  2. Install Required Packages:

    pip install -r requirements.txt

  3. Install ``pytagi`` from PyPI:

    pip install pytagi

  4. Test the Installation:

    python -m examples.classification

    Note

    You can develop your own applications using the provided examples (see python_examples).

Local Installation#

  1. Clone the Repository:

    git clone https://github.com/lhnguyen102/cuTAGI.git
cd cuTAGI
git submodule update --init --recursive

    Note

    The git submodule command clones the pybind11 repository, required for binding Python with C++/CUDA.

  2. Ensure Conda Environment is Activated.

    conda activate your_env_name

  3. Install Required Packages:

    pip install -r requirements.txt

  4. Install pytagi Locally:

    Remove cache

    pip cache purge

    Install package

    pip install .

  5. Test the Installation:

    build/run_tests

Installing cuTAGI#

cuTAGI is the native C++/CUDA implementation. Using Docker is the recommended method for installation.

Docker Build#

  1. Install Docker: Follow the official instructions.

  2. Ensure CUDA Compatibility:

    • Ensure the host machine’s CUDA version is compatible with the Docker image’s CUDA version (>=12.2).

    • Install the NVIDIA Container Toolkit to enable Docker to use the NVIDIA GPU:

      dpkg -l | grep nvidia-container-toolkit

  3. Build the Docker Image:

    • CPU Build:

      scripts/docker_build.sh

    • CUDA Build:

      scripts/docker_build.sh device=cuda version=latest

  4. Run Unit Tests:

    • For CPU:

      scripts/docker_run.sh

    • For CUDA (GPU):

      scripts/docker_run.sh device=cuda version=latest

    Note

    Ensure Docker is running during the build and run processes.

Ubuntu 22.04 Installation (Without Docker)#

  1. Install CUDA Toolkit:

    • Download and install the CUDA Toolkit (version >= 12.2).

    • Add the following to your ~/.bashrc file to update your PATH:

      export PATH="/usr/local/cuda-12.2/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda-12.2/lib64:$LD_LIBRARY_PATH"

  2. Install GCC Compiler:

    sudo apt install g++

  3. Install CMake:

    Follow the official instructions to install CMake.

  4. Build the Project:

    Navigate to the cuTAGI folder and run:

    scripts/compile.sh [option]

    Replace [option] with one of the following:

    • Release: For optimized release build (Default)

    • ReleaseWithInfo: For release build with debug information

    • Debug: For debug build using GDB Debugger

macOS (CPU Version)#

cuTAGI supports CPU-only builds on macOS.

  1. Install Xcode.

  2. Install CMake:

    Refer to the Installation Guide: CMake on macOS section below.

  3. Build the Project:

    Navigate to the cuTAGI folder and run:

    scripts/compile.sh [option]

Running Unit Tests#

  • For C++:

    build/run_tests

  • For Python:

    python -m test.py_unit.main

Installation Guide: CMake on macOS#

Step 1: Uninstall Previous CMake Versions (Optional)#

sudo find /usr/local/bin -type l -lname '/Applications/CMake.app/*' -delete
sudo rm -rf /Applications/CMake.app

Step 2: Install CMake#

  1. Download the installer:

    mkdir ~/Downloads/CMake
curl --silent --location --retry 3 "https://github.com/Kitware/CMake/releases/download/v3.30.5/cmake-3.30.5-macos-universal.dmg" --output ~/Downloads/CMake/cmake-macos.dmg

  2. Mount the image:

    yes | PAGER=cat hdiutil attach -quiet -mountpoint /Volumes/cmake-macos ~/Downloads/CMake/cmake-macos.dmg

  3. Copy CMake app to Applications:

    cp -R /Volumes/cmake-macos/CMake.app /Applications/

  4. Unmount the image:

    hdiutil detach /Volumes/cmake-macos

  5. Add CMake to the PATH:

    sudo "/Applications/CMake.app/Contents/bin/cmake-gui" --install=/usr/local/bin

  6. Verify the installation:

    cmake --version

  7. Clean up:

    rm -rf ~/Downloads/CMake
Contents