数据与计算发展前沿 ›› 2020, Vol. 2 ›› Issue (1): 27-37.doi: 10.11871/jfdc.issn.2096-742X.2020.01.003

• 专刊:高性能与高通量计算及应用 • 上一篇    下一篇

面向人工智能和大数据的高效能计算

李肯立1,2,*(),阳王东1,*(),陈岑1,2,陈建国1,丁岩1   

  1. 1.湖南大学信息科学与工程学院,湖南 长沙 410008
    2.国家超级计算长沙中心,湖南 长沙 410008
  • 收稿日期:2019-10-28 出版日期:2020-02-20 发布日期:2020-03-28
  • 通讯作者: 李肯立,阳王东 E-mail:lkl@hnu.edu.cn;yangwangdong@163.com
  • 作者简介:李肯立,湖南大学信息科学与工程学院,教授,博士生导师,主要研究方向为高性能计算、并行计算、人工智能。
    本文承担工作为:框架的整体结构设计、研究指导。
    Li Kenli, Doctor, is a professor of School of Information Science and Engineering, Hunan University. His main research fields are high performance computing, parallel computing and artificial intelligence.
    He undertakes the following tasks: the overall structure design and research guidance of the frame.|阳王东,湖南大学信息科学与工程学院,教授,博士生导师,主要研究方向为高性能计算、并行计算。
    本文承担工作为:研究方向的凝练和论文的整合。
    Yang Wangdong, Doctor, is a professor of School of Information Science and Engineering, Hunan University. His main research fields are high performance computing, parallel computing.
    He undertakes the following tasks: the figure research direction out and the integration of papers.|陈岑,湖南大学信息科学与工程学院博士后,主要研究方向为大数据处理、并行计算与人工智能。
    本文承担工作为:序言撰写和研究问题分析。
    Chen Cen, post-doctoral researcher at the School of Information Science and Engineering, Hunan University, focuses on big data processing, parallel computing and artificial intelligence.
    He undertakes the following tasks: preface writing and problem analysis.
    E-mail: chencen@hnu.edu.cn|陈建国,湖南大学信息科学与工程学院,博士后,主要研究方向为大数据和人工智能。
    本文承担工作为:框架的整体结构设计、研究指导。面向大数据和人工智能的高效能计算所面临的挑战分析。
    Chen Jianguo, is a post-doctoral researcher at School of Information Science and Engineering, Hunan University. His major research areas include big data and artificial intelligence. He undertakes the following tasks: being the research director who is responsible for the design of the whole framework and analyzing the challenges of efficient computing for big data and artificial intelligence.
    E-mail: jianguochen@hnu.edu.cn|丁岩,湖南大学信息科学与工程学院在读博士生,主要研究方向为边缘计算、数据挖掘。
    本文承担工作为:深度神经网络模型剪枝与压缩方法调研。
    Ding Yan, a PhD student at College of Information Science and Engineering, Hunan University. His research fields are edge computing and data mining.
    He undertakes the following tasks: investigate on pruning and compression methods of deep neural network models.
    E-mail: ding@hnu.edu.cn
  • 基金资助:
    国家重点研发计划(2018YFB1003401);国家杰出青年基金项目(61625202);国家自然科学基金项目(61572175);国家自然科学基金项目(61572175);国家自然科学基金项目(61751204);国家自然科学基金项目(61472124);国际交流合作项目(61860206011)

Efficient Computing for Artificial Intelligence and Big Data

Li kenli1,2,*(),Yang Wangdong1,*(),Chen Cen1,2,Chen Jianguo1,Ding Yan1   

  1. 1.College of Information Science and Engineering, Hunan University, Changsha ,Hunan 410008, China
    2.National Super-computer Center in Changsha, Changsha ,Hunan 410008, China
  • Received:2019-10-28 Online:2020-02-20 Published:2020-03-28
  • Contact: Li kenli,Yang Wangdong E-mail:lkl@hnu.edu.cn;yangwangdong@163.com

摘要:

[目的]本文主要分析人工智能和大数据应用随着迅速增大的数据规模,给计算机系统带来的主要挑战,并针对计算机系统的发展趋势给出了一些面向人工智能和大数据亟待解决的高效能计算的若干研究方向。[文献范围]本文广泛查阅国内外在超级计算和高性能计算平台进行大数据和人工智能计算的最新研究成果及解决的挑战性问题。[方法]大数据既为人工智能提供了日益丰富的训练数据集合,但也给计算机系统的算力提出了更高的要求。近年来我国超级计算机处于世界的前列,为大数据和人工智能的大规模应用提供了强有力的计算平台支撑。[结果]而目前以超级计算机为代表的高性能计算平台大多采用CPU+加速器构成的异构并行计算系统,其数量众多的计算核心能够为人工智能和大数据应用提供强大的计算能力。[局限性]由于体系结构复杂,在充分发挥计算能力和提高计算效率方面存在较大挑战。尤其针对有别于科学计算的人工智能和大数据领域,其并行计算效率的提升更为困难。[结论]因此需要从底层的资源管理、任务调度、以及基础算法设计、通信优化,到上层的模型并行化和并行编程等方面展开高效能计算的研究,全面提升人工智能和大数据应用在高性能计算平台上的计算能效。

关键词: 超级计算, 大数据, 高效能计算, 人工智能, 异构系统

Abstract:

[Objective] This paper mainly analyses the main challenges brought to computer system by the rapid increase of data scale of AI and big data application. In view of the development trend of computer system, some research directions of high-efficiency computing towards AI and big data are given. [Coverage] In this paper, the latest research results and challenges of big data and artificial intelligence computing on supercomputing and high performance computing platforms at home and abroad are extensively surveyed. [Methods] Big data not only provides an increasingly rich training data set for artificial intelligence, but also puts forward higher requirements for the computing power of computer systems. In recent years, China's supercomputer techniques are at the forefront of the world, which provides a powerful computing platform for large-scale applications of big data and artificial intelligence. [Results] At present, high-performance computing platforms represented by supercomputers mostly use heterogeneous parallel computing systems composed of CPUs and accelerators, where a large number of computing cores can provide powerful computing power for AI and big data applications. [Limitations]However, due to the complex architecture, there are major challenges in making full use of computing power and improving computing efficiency. The parallel computing efficiency is more difficult to improve, especially in the artificial intelligence and big data domains which are different from scientific computing. [Conclusions] Therefore, it is required to conduct research on high-performance computing from underlying resource management, task scheduling, basic algorithm design, and communication optimization to the upper level of model parallelization, so that the computational efficiency of artificial intelligence and big data applications on high-performance computing platforms can be improved.

Key words: artificial intelligence, big data, heterogeneous systems, high efficiency computing, supercomputing