Shiyang Li

Shiyang Li

PhD student at University of Minnesota-Twin Cities

OneGraph: A Cross-Architecture Framework for Large-Scale Graph Computing on GPUs Based on oneAPI

Published in CCF Transactions on High Performance Computing (CCF-THPC), 2024

Abstract

We present OneGraph, a cross-architecture graph computing framework built on Intel’s oneAPI that enables portable and efficient large-scale graph processing across different GPU architectures (NVIDIA, AMD, Intel). By abstracting away hardware-specific details through SYCL, OneGraph achieves near-native performance while maintaining code portability.

Key Contributions

  • A cross-architecture graph computing framework based on oneAPI/SYCL
  • Portable graph processing across NVIDIA, AMD, and Intel GPUs
  • Performance optimizations that achieve near-native efficiency on multiple GPU architectures

Authors

Shiyang Li, Jingyu Zhu, Jiaxun Han, Yuting Peng, Zhuoran Wang, Xiaoli Gong, Gang Wang, Jin Zhang, Xuqiang Wang

CCF Transactions on High Performance Computing (CCF-THPC), 6(2): 179-191, 2024.

OneGraph Architecture

Recommended citation: S. Li, J. Zhu, J. Han, Y. Peng, Z. Wang, X. Gong, G. Wang, J. Zhang, X. Wang. "OneGraph: A Cross-Architecture Framework for Large-Scale Graph Computing on GPUs Based on oneAPI." CCF Transactions on High Performance Computing (CCF-THPC), 6(2): 179-191, 2024.
Download Paper