一句话：提出NDPipe，在照片存储服务器中借助近数据处理与商品GPU，实现可扩展的DNN推理和持续训练。
NDPipe面向照片存储场景中图像数据持续增长而推理与训练需频繁访问数据的问题，提出将计算下沉到存储服务器。系统在数据中心内为存储服务器配置低成本商品GPU，利用其集体算力在图像数据附近执行DNN推理与训练；通过高效划分DNN模型并挖掘多存储服务器的数据并行性，降低同步开销、提升训练吞吐，并优化近数据处理引擎以充分利用各存储服务器的系统组件。评估表明，在相同能耗预算下，NDPipe的推理吞吐量比典型照片存储系统高1.39倍，训练速度快2.64倍。
This paper proposes a novel photo storage system called NDPipe, which accelerates the performance of training and inference for image data by leveraging near-data processing in photo storage servers. NDPipe distributes storage servers with inexpensive commodity GPUs in a data center and uses their collective intelligence to perform inference and training near image data. By efficiently partitioning deep neural network (DNN) models and exploiting the data parallelism of many storage servers, NDPipe can achieve high training throughput with low synchronization costs. NDPipe optimizes the near-data processing engine to maximally utilize system components in each storage server. Our results show that, given the same energy budget, NDPipe exhibits 1.39× higher inference throughput and 2.64× faster training speed than typical photo storage systems.
@inproceedings{KimOKK024,
title = {NDPipe: Exploiting Near-data Processing for Scalable Inference and Continuous Training in Photo Storage},
author = {Jungwoo Kim and Seonggyun Oh and Jaeha Kung and Yeseong Kim and Sungjin Lee},
booktitle = {ASPLOS},
year = {2024},
doi = {10.1145/3620666.3651345}
}