Everest: GPU-Accelerated System For Mining Temporal Motifs

被引:1
作者
Yuan, Yichao [1 ]
Ye, Haojie [1 ]
Vedula, Sanketh [2 ]
Kaza, Wynn [1 ]
Talati, Nishil [1 ]
机构
[1] Univ Michigan, Ann Arbor, MI 48109 USA
[2] Technion, Haifa, Israel
来源
PROCEEDINGS OF THE VLDB ENDOWMENT | 2023年 / 17卷 / 02期
关键词
NETWORK MOTIFS; EFFICIENT;
D O I
10.14778/3626292.3626299
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Temporal motif mining is the task of finding the occurrences of subgraph patterns within a large input temporal graph that obey the specified structural and temporal constraints. Despite its utility in several critical application domains that demand high performance (e.g., detecting fraud in financial transaction graphs), the performance of existing software is limited on commercial hardware platforms, in that it runs for tens of hours. This paper presents Everest-a system that efficiently maps the workload of mining (supports both enumeration and counting) temporal motifs to the highly parallel GPU architecture. In particular, using an input temporal graph and a more expressive user-defined temporal motif query definition compared to prior works, Everest generates an execution plan and runtime primitives that optimize the workload execution by exploiting the high compute throughput of a GPU. Everest generates motif-specific mining code to reduce long-latency memory accesses and frequent thread divergence operations. Everest incorporates novel low-cost runtime mechanisms to enable load balancing to improve GPU hardware utilization. To support large graphs that do not fit on GPU memory, Everest also supports multi-GPU execution by intelligently partitioning the edge list that prevents inter-GPU communication. Everest hides the implementation complexity of presented optimizations away from the targeted system user for better usability. Our evaluation shows that, using proposed optimizations, Everest improves the performance of a baseline GPU implementation by 19x, on average.
引用
收藏
页码:162 / 174
页数:13
相关论文
共 40 条
[1]   Network motifs: theory and experimental approaches [J].
Alon, Uri .
NATURE REVIEWS GENETICS, 2007, 8 (06) :450-461
[2]  
[Anonymous], 2023, Unified Memory for CUDA Beginners
[3]   Improving Graph Neural Network Expressivity via Subgraph Isomorphism Counting [J].
Bouritsas, Giorgos ;
Frasca, Fabrizio ;
Zafeiriou, Stefanos ;
Bronstein, Michael M. .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (01) :657-668
[4]  
Chamberlain BP, 2022, Arxiv, DOI arXiv:2209.15486
[5]  
Chen XH, 2022, PROCEEDINGS OF THE 16TH USENIX SYMPOSIUM ON OPERATING SYSTEMS DESIGN AND IMPLEMENTATION, OSDI 2022, P857
[6]   Pangolin: An Efficient and Flexible Graph Mining System on CPU and GPU [J].
Chen, Xuhao ;
Dathathri, Roshan ;
Gill, Gurbinder ;
Pingali, Keshav .
PROCEEDINGS OF THE VLDB ENDOWMENT, 2020, 13 (08) :1190-1205
[7]   Scalable Motif Counting for Large-scale Temporal Graphs [J].
Gao, Zhongqiang ;
Cheng, Chuanqi ;
Yu, Yanwei ;
Cao, Lei ;
Huang, Chao ;
Dong, Junyu .
2022 IEEE 38TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE 2022), 2022, :2656-2668
[8]   Bridging the Gap: A Pragmatic Approach to Generating Insider Threat Data [J].
Glasser, Joshua ;
Lindauer, Brian .
IEEE CS SECURITY AND PRIVACY WORKSHOPS (SPW 2013), 2013, :98-104
[9]  
Hajdu L, 2020, COMM COM INF SC, V1260, P145, DOI 10.1007/978-3-030-55814-7_12
[10]   A Deeper Dive into Pattern-Aware Subgraph Exploration with PEREGRINE [J].
Jamshidi K. ;
Vora K. .
Operating Systems Review (ACM), 2021, 55 (01) :1-10