TY - GEN
T1 - Understanding and bridging the gaps in current GNN performance optimizations
AU - Huang, Kezhao
AU - Zhai, Jidong
AU - Zheng, Zhen
AU - Yi, Youngmin
AU - Shen, Xipeng
N1 - Publisher Copyright:
© 2021 ACM.
PY - 2021/2/17
Y1 - 2021/2/17
N2 - 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 state-of-the-art GPU optimization techniques and tailor them to the special properties of GNN. Experimental results show that these optimizations achieve 1.37× - 15.5× performance improvement over the state-of-the-art frameworks on various GNN models.
AB - 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 state-of-the-art GPU optimization techniques and tailor them to the special properties of GNN. Experimental results show that these optimizations achieve 1.37× - 15.5× performance improvement over the state-of-the-art frameworks on various GNN models.
KW - GNN
KW - parallelism
KW - performance optimizations
UR - http://www.scopus.com/inward/record.url?scp=85101688246&partnerID=8YFLogxK
U2 - 10.1145/3437801.3441585
DO - 10.1145/3437801.3441585
M3 - Conference contribution
AN - SCOPUS:85101688246
T3 - Proceedings of the ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPOPP
SP - 119
EP - 132
BT - PPoPP 2021 - Proceedings of the 2021 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming
PB - Association for Computing Machinery
T2 - 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, PPoPP 2021
Y2 - 27 February 2021 through 3 March 2021
ER -