GraphOne+, a unified heterogeneous platform that supports graph analysis problems including graph computation, graph learning, and graph mining. This platform aims to provide users with a simple interface to implement graph algorithms and efficiently perform computations on heterogeneous devices. Currently, it supports the computation of graph computation problems on GPU devices.

• Operating System: Linux (Ubuntu 18.04 and above)
• Hardware: NVIDIA GPU (preferably with a Compute Capability of 7.0 or higher)
• Software:
  • GCC (version 9.4.0 and above)
  • CUDA (version 12.4 and above)
  • CMake (version 3.8 and above)
  • LibTorch (version 2.1.0 and above)

Execute the following instructions in the root directory of the project. The generated executable files will be located in the build/apps directory.

mkdir build && cd build
cmake .. -DCMAKE_PREFIX_PATH=<your_libtorch_path>
make -j

Support for matrix market format graph datasets.
You can download graph datasets from the website https://ogb.stanford.edu/.

You can directly run executable programs located in the build/apps directory after building the project.

./apps/sssp -i ../datasets/ak2010/ak2010.adj --src 0

GraphOne is developed at SCTS&CGCL Lab (http://grid.hust.edu.cn/) by Haoyan Huang, Yu Huang, and Long Zheng. For any questions, please contact Haoyan Huang ([email protected]), Yu Huang ([email protected]) and Long Zheng ([email protected]).