Write a Blog >>
PPoPP 2021
Sat 27 February - Wed 3 March 2021
Mon 1 Mar 2021 12:30 - 12:45 - Session 3. Graphs Chair(s): Ang Li

Graph Neural Network (GNN) has recently drawn a rapid increase of interest in many domains for its effectiveness in learning over graphs. Maximizing its performance is essential for many tasks, but remains preliminarily understood. In this work, we provide an in-depth examination of the state-of-the-art GNN frameworks, revealing five major gaps in the current frameworks in optimizing GNN performance, especially in handling the special complexities of GNN over traditional graph or DNN operations. Based on the insights, we put together a set of optimizations to fill the gaps. These optimizations leverage the start-of-the-art GPU optimization techniques and tailor them to the special properties of GNN. Experimental results show that these optimizations achieve 1.37–15.5X performance improvement over the state-of-the-art frameworks on various GNN models.

Mon 1 Mar
Times are displayed in time zone: Eastern Time (US & Canada) change

12:30 - 13:30
Session 3. GraphsMain Conference
Chair(s): Ang LiPacific Northwest National Laboratory
12:30
15m
Talk
Understanding and Bridging the Gaps in Current GNN Performance Optimizations
Main Conference
Kezhao HuangTsinghua University, Jidong ZhaiTsinghua University, Zhen ZhengAlibaba Group, Youngmin YiUniversity of Seoul, Xipeng ShenNorth Carolina State University
Link to publication
12:45
15m
Talk
A Fast Work-Efficient SSSP Algorithm for GPUs
Main Conference
Kai WangUniversity of Texas at Austin, Donald FussellUniversity of Texas at Austin, Calvin LinUniversity of Texas at Austin
Link to publication
13:00
15m
Talk
ShadowVM: Accelerating Data Plane for Data Analytics with Bare Metal CPUs and GPUs
Main Conference
Zhifang LiEast China Normal University, Mingcong HanEast China Normal University, Shangwei WuEast China Normal University, Chuliang WengEast China Normal University
Link to publication
13:15
15m
Talk
BiPart: A Parallel and Deterministic Hypergraph Partitioner
Main Conference
Sepideh MalekiThe University of Texas at Austin, Udit AgarwalUT Austin, Martin BurtscherTexas State University, Keshav PingaliThe University of Texas at Austin
Link to publication