LightRW: FPGA Accelerated Graph Dynamic Random Walks

Author:

Tan Hongshi1ORCID,Chen Xinyu1ORCID,Chen Yao1ORCID,He Bingsheng1ORCID,Wong Weng-Fai1ORCID

Affiliation:

1. National University of Singapore, Singapore, Singapore

Abstract

Graph dynamic random walks (GDRWs) have recently emerged as a powerful paradigm for graph analytics and learning applications, including graph embedding and graph neural networks. Despite the fact that many existing studies optimize the performance of GDRWs on multi-core CPUs, massive random memory accesses and costly synchronizations cause severe resource underutilization, and the processing of GDRWs is usually the key performance bottleneck in many graph applications. This paper studies an alternative architecture, FPGA, to address these issues in GDRWs, as FPGA has the ability of hardware customization so that we are able to explore fine-grained pipeline execution and specialized memory access optimizations. Specifically, we propose LightRW, a novel FPGA-based accelerator for GDRWs. LightRW embraces a series of optimizations to enable fine-grained pipeline execution on the chip and to exploit the massive parallelism of FPGA while significantly reducing memory accesses. As current commonly used sampling methods in GDRWs do not efficiently support fine-grained pipeline execution, we develop a parallelized reservoir sampling method to sample multiple vertices per cycle for efficient pipeline execution. To address the random memory access issues, we propose a degree-aware configurable caching method that buffers hot vertices on-chip to alleviate random memory accesses and a dynamic burst access engine that efficiently retrieves neighbors. Experimental results show that our optimization techniques are able to improve the performance of GDRWs on FPGA significantly. Moreover, LightRW delivers up to 9.55x and 9.10x speedup over the state-of-the-art CPU-based MetaPath and Node2vec random walks, respectively. This work is open-sourced on GitHub at https://github.com/Xtra-Computing/LightRW.

Funder

AI Singapore Programme

Google South \& Southeast Asia Research Award 2022

AMD Heterogeneous Accelerated Compute Clusters (HACC) program

Ministry of Education AcRF Tier 2 grant

Publisher

Association for Computing Machinery (ACM)

Reference68 articles.

1. AMD. 2023. Heterogeneous Accelerated Compute Clusters (HACC) Program. https://www.amd-haccs.io/index.html. AMD. 2023. Heterogeneous Accelerated Compute Clusters (HACC) Program. https://www.amd-haccs.io/index.html.

2. Cascade Lake: Next Generation Intel Xeon Scalable Processor

3. Rabbit Order: Just-in-Time Parallel Reordering for Fast Graph Analysis

4. DynaBurst: Dynamically Assemblying DRAM Bursts over a Multitude of Random Accesses

5. Survey on hardware implementation of random number generators on FPGA: Theory and experimental analyses

Cited by 4 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. An FPGA-Based Accelerator for Graph Embedding using Sequential Training Algorithm;2024 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW);2024-05-27

2. Observations and Opportunities in Solving Large-Scale Graph Data Processing Challenges at ByteDance by Using Heterogeneous Hardware;2024 IEEE 40th International Conference on Data Engineering (ICDE);2024-05-13

3. F-TADOC: FPGA-Based Text Analytics Directly on Compression with HLS;2024 IEEE 40th International Conference on Data Engineering (ICDE);2024-05-13

4. FlowWalker: A Memory-Efficient and High-Performance GPU-Based Dynamic Graph Random Walk Framework;Proceedings of the VLDB Endowment;2024-04

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3