HiHGNN: Accelerating HGNNs Through Parallelism and Data Reusability Exploitation
IEEE Transactions on Parallel and Distributed Systems · 2024-07 · 35 (7) · 1122-1138
DOI: 10.1109/tpds.2024.3394841 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Parallel and Distributed Systems, 2024-07. Verfasst von Runzhen Xue und 10 weiteren Autor:innen. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 79 Literaturangaben. Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, CAS Project for Young Scientists in Basic Research.
Abstract (Originalsprache)
Heterogeneous graph neural networks (HGNNs) have emerged as powerful algorithms for processing heterogeneous graphs (HetGs), widely used in many critical fields. To capture both structural and semantic information in HetGs, HGNNs first aggregate the neighboring feature vectors for each vertex in each semantic graph and then fuse the aggregated results across all semantic graphs for each vertex. Unfortunately, existing graph neural network accelerators are ill-suited to accelerate HGNNs. This is because they fail to efficiently tackle the specific execution patterns and exploit the high-degree parallelism as well as data reusability inside and across the processing of semantic graphs in HGNNs. In this work, we first quantitatively characterize a set of representative HGNN models on GPU to disclose the execution bound of each stage, inter-semantic-graph parallelism, and inter-semantic-graph data reusability in HGNNs. Guided by our findings, we propose a high-performance HGNN accelerator, HiHGNN, to alleviate the execution bound and exploit the newfound parallelism and data reusability in HGNNs. Specifically, we first propose a bound-aware stage-fusion methodology that tailors to HGNN acceleration, to fuse and pipeline the execution stages being aware of their execution bounds. Second, we design an independency-aware parallel execution design to exploit the inter-semantic-graph parallelism. Finally, we present a similarity-aware execution scheduling to exploit the inter-semantic-graph data reusability. Compared to the state-of-the-art software framework running on NVIDIA GPU T4 and GPU A100, HiHGNN respectively achieves an average 40.0× and 8.3× speedup as well as 99.59% and 99.74% energy reduction with quintile the memory bandwidth of GPU A100.
Förderung
- National Key Research and Development Program of China
- National Natural Science Foundation of China
- CAS Project for Young Scientists in Basic Research
- CAS Project for Youth Innovation Promotion Association
- Beijing Natural Science Foundation
- Beijing Nova Program
Zitieren
APA
Xue, R., Han, D., Yan, M., Zou, M., Yang, X., Wang, D., Li, W., Tang, Z., Kim, J., Ye, X., & Fan, D. (2024). HiHGNN: Accelerating HGNNs Through Parallelism and Data Reusability Exploitation. IEEE Transactions on Parallel and Distributed Systems, 35(7), 1122-1138. https://doi.org/10.1109/tpds.2024.3394841
Vancouver
Xue R, Han D, Yan M, Zou M, Yang X, Wang D, et al. HiHGNN: Accelerating HGNNs Through Parallelism and Data Reusability Exploitation. IEEE Transactions on Parallel and Distributed Systems. 2024;35(7):1122-1138. doi:10.1109/tpds.2024.3394841
BibTeX
@article{Xue2024,
author = {Xue, Runzhen and Han, Dengke and Yan, Mingyu and Zou, Mo and Yang, Xiaocheng and Wang, Duo and Li, Wenming and Tang, Zhimin and Kim, John and Ye, Xiaochun and Fan, Dongrui},
title = {HiHGNN: Accelerating HGNNs Through Parallelism and Data Reusability Exploitation},
journal = {IEEE Transactions on Parallel and Distributed Systems},
year = {2024},
volume = {35},
number = {7},
pages = {1122--1138},
doi = {10.1109/tpds.2024.3394841}
}
EUVIMED