Dynamic Multi-Resolution Data Storage

Hu, Yu-Ching; Lokhandwala, Murtuza; I, Te; Tseng, Hung-Wei

doi:10.1145/3352460.3358282

Cited by 8 publications

(1 citation statement)

References 76 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The focus of our work is on accelerating large-scale GNN training with an ISP architecture. There is a large body of prior literature exploring in-storage/near-data processing [1], [5], [10], [12], [13], [16], [25], [27], [31]- [33], [35], [38], [38], [39], [42], [49], [51], [54], [58], [65], [67], [72]- [74], [76], [81], [81]- [83] or in-memory processing [2]- [4], [9], [18], [20], [23], [29], [34], [36], [37], [44], [45], [50], [60], [68], [79] architectures for data-intensive workloads as well as ASIC/FPGA/GPU based acceleration for graph neural networks [24], [40], [52], [54]-…”

Section: Related Workmentioning

confidence: 99%

SmartSAGE: Training Large-scale Graph Neural Networks using In-Storage Processing Architectures

Lee¹,

Chung²,

Rhu³

2022

Preprint

View full text Add to dashboard Cite

Graph neural networks (GNNs) can extract features by learning both the representation of each objects (i.e., graph nodes) and the relationship across different objects (i.e., the edges that connect nodes), achieving state-of-the-art performance in various graph-based tasks. Despite its strengths, utilizing these algorithms in a production environment faces several challenges as the number of graph nodes and edges amount to several billions to hundreds of billions scale, requiring substantial storage space for training. Unfortunately, state-of-the-art ML frameworks employ an in-memory processing model which significantly hampers the productivity of ML practitioners as it mandates the overall working set to fit within DRAM capacity. In this work, we first conduct a detailed characterization on a state-of-theart, large-scale GNN training algorithm, GraphSAGE. Based on the characterization, we then explore the feasibility of utilizing capacity-optimized NVMe SSDs for storing memory-hungry GNN data, which enables large-scale GNN training beyond the limits of main memory size. Given the large performance gap between DRAM and SSD, however, blindly utilizing SSDs as a direct substitute for DRAM leads to significant performance loss. We therefore develop SmartSAGE, our software/hardware co-design based on an in-storage processing (ISP) architecture. Our work demonstrates that an ISP based large-scale GNN training system can achieve both high capacity storage and high performance, opening up opportunities for ML practitioners to train large GNN datasets without being hampered by the physical limitations of main memory size.

show abstract