Aidge Backend CUDA

Aidge Backend CUDA is the implementation library for CUDA-based operators within the Aidge framework.

Quick Start

Prerequisite:

• aidge_core

pip install aidge-backend-cuda

[!WARNING] It is highly recommended to build this module from source rather than using the pypi package.

🛠 Build from Source

Prerequisite (in addition to previous one):

• Please review the global installation instructions before proceeding.
• If using a virtual environment, use the same one for all Aidge modules.

• CUDA ToolKit 12.6
  • 12.4.1 is the strict minimum, but 12.6+ is recommended.
• CUDnn9
  • make sure to install the CUDA 12 compatible version

[!TIP] On Windows at least, CUDnn9 headers and libraries needs to be findable. One way is to add files from CUDnn to CUDA in their respective bin/, lib/ and include/. Or you could add these folders to your path.

1. Python or C++ installation using setup scripts

[!TIP] Use Get-Help setup.ps1 (Windows) or ./setup.sh -h (Unix) for full documentation.

2. Python Installation using pip

Run these commands from the aidge_backend_cuda/ directory:

# Standard install
pip install . -v

# Install with testing dependencies
pip install .[test] -v && pytest

Editable Install (Experimental)

Use this for real-time development without re-installing.

pip install --no-build-isolation -ve . --config-settings=editable.rebuild=true -Cbuild-dir=build

3. C++ Installation (CMake)

A CMakePresets.json is provided for standard configurations.

# Configure, Build, and Install
cmake --preset clang-debug
cmake --build --preset clang-debug
cmake --install

# Run C++ Tests
ctest --test-dir build/

[!TIP] Create a CMakeUserPresets.json to define your own local build configurations.