arXiv 2103.080532021-03-14cs.DCcs.SI

TRUST：在GPU上重新设计的三角形计数

TRUST: Triangle Counting Reloaded on GPUs

Stevens Institute of Technology（史蒂文斯理工学院）；State Key Laboratory for Novel Software Technology, Nanjing University（南京大学软件新技术国家重点实验室）；Huazhong University of Science and Technology（华中科技大学）；Computational Research Division, Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室计算研究部）；Brookhaven National Laboratory（布鲁克海文国家实验室）；Department of Computer Science & Engineering, University of Connecticut（康涅狄格大学计算机科学与工程系）；Department of Electrical Engineering and Computer Science, University of California, Merced（加州大学默塞德分校电气与计算机工程系）

Santosh Pandey, Zhibin Wang, Sheng Zhong, Chen Tian, Bolong Zheng, Xiaoye Li, Lingda Li, Adolfy Hoisie, Caiwen Ding, Dong Li, Hang Liu

DOI：: 10.1109/TPDS.2021.3064892

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

TRUST通过哈希操作和顶点中心机制实现大规模GPU三角形计数，首次达到每秒超过一万亿遍历边数（TEPS）的性能。

摘要：

三角计数是许多图应用的基础构建块。传统观点认为i）哈希不适合三角计数，ii）边中心的三角计数优于顶点中心的设计，iii）通信自由且工作负载平衡的图划分是三角计数的巨大挑战。相反，我们主张i）哈希可以帮助在图形处理单元（GPU）上可扩展的三角计数的关键操作，即列表交集和图划分，ii）顶点中心设计减少了哈希表构造成本和内存消耗，这在GPU上受到限制。此外，iii）我们利用图和工作负载协同，并基于哈希的2D划分来扩展顶点中心的三角计数，使其能够在1000多台GPU上持续扩展。在本工作中，我们提出了TRUST，其核心是哈希操作和顶点中心机制的三角计数。到目前为止，TRUST是首个达到每秒超过一万亿遍历边数（TEPS）速率的三角计数工作。

英文摘要：

Triangle counting is a building block for a wide range of graph applications. Traditional wisdom suggests that i) hashing is not suitable for triangle counting, ii) edge-centric triangle counting beats vertex-centric design, and iii) communication-free and workload balanced graph partitioning is a grand challenge for triangle counting. On the contrary, we advocate that i) hashing can help the key operations for scalable triangle counting on Graphics Processing Units (GPUs), i.e., list intersection and graph partitioning, ii)vertex-centric design reduces both hash table construction cost and memory consumption, which is limited on GPUs. In addition, iii) we exploit graph and workload collaborative, and hashing-based 2D partitioning to scale vertex-centric triangle counting over 1,000 GPUswith sustained scalability. In this work, we present TRUST which performs triangle counting with the hash operation and vertex-centric mechanism at the core. To the best of our knowledge, TRUSTis the first work that achieves over one trillion Traversed Edges Per Second (TEPS) rate for triangle counting.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。