OSDI '23 - MGG: Accelerating Graph Neural Networks with Fine-Grained Intra-Kernel Communication...

OSDI '23 - MGG: Accelerating Graph Neural Networks with Fine-Grained Intra-Kernel Communication-Computation Pipelining on Multi-GPU Platforms

Yuke Wang, University of California Santa Barbara, Boyuan Feng, University of California Santa Barbara, Zheng Wang, University of California Santa Barbara, Tong Geng, University of Rochester, Kevin Barker, Pacific Northwest National Laboratory, Ang Li, Pacific Northwest National Laboratory, Yufei Ding, University of California Santa Barbara

The increasing size of input graphs for graph neural networks (GNNs) highlights the demand for using multi-GPU platforms. However, …

california communication computation gpu graph kernel kevin national national laboratory networks neural networks pacific pacific northwest pacific northwest national laboratory platforms university university of california wang