Challenges and suggestions for development of intelligent computing in China

doi:10.12267/j.issn.2096-5931.2024.11.013

Abstract

Abstract:

Firstly, this article analyzes the three typical stages of the development of computing power centers in China in recent years, including traditional data centers, customized data centers, and intelligent computing centers. Subsequently, five challenges faced by the development of intelligent computing are studied, including chips and networks, computing facilities, software ecology, computing power utilization, and energy demand. Finally, suggestions are put forward for the development of China’s intelligent computing industry, including technical research, standard formulation, system compatibility, computing and network convergence, and international cooperation.

Key words: intelligent computing, computing power center, computing and network convergence, international cooperation

CLC Number:

TP30
F124

GUO Liang, LIU Rui. Challenges and suggestions for development of intelligent computing in China[J]. Information and Communications Technology and Policy, 2024, 50(11): 85-89.

Add to citation manager EndNote|Ris|BibTeX

URL:

http://ictp.caict.ac.cn/EN/10.12267/j.issn.2096-5931.2024.11.013

http://ictp.caict.ac.cn/EN/Y2024/V50/I11/85

References 14

[1]	JIANG Z H, LIN H B, ZHONG Y M, et al. MegaScale: scaling large language model training to more than 10 000 GPUs[J]. arXiv Preprint. arXiv:2402.15627, 2024.
[2]	中国信息通信研究院. 2023智能算力发展白皮书[R], 2023.
[3]	中关村在线. 哪些地方适合建智算中心?[EB/OL]. (2024-10-04)[2024-10-09]. https://news.qq.com/rain/a/20241004A02MO700.
[4]	工业和信息化部. 2010年全国电信业统计公报[R], 2011.
[5]	华为技术有限公司. 计算2030[R], 2024.
[6]	AVELAR V, AZEVEDO D, FRENCH A. WP49-PUE a comprehensive examination of the metric[R], 2012.
[7]	NVIDIA. NVIDIA blackwell architecture technical brief[R], 2024.
[8]	郑纬民, 翟季冬, 翟明书. 智能算力核心基础系统软件的现状与展望[J]. 中兴通讯技术, 2024, 30(2):2-8.
[9]	DONG J B, WANG S C, FENG F, et al. ACCL: architecting highlyscalable distributed training systems with highly efficient collective communication library[J]. IEEE Micro, 2021, 41(5):85-92. DOI: 10.1109/MM.2021.3091475.
[10]	华为技术有限公司. 昇思通信[R], 2020.
[11]	中国信息通信研究院. 发布!我国算力中心相关数据[EB/OL]. (2024-06-28)[2024-10-09]. https://mp.weixin.qq.com/s/kjQUNtlv_Y7QL4spjhk0DQ.
[12]	人民邮电报. 网信办备案AI大模型数量已达197行业大模型占比近70%[EB/OL]. (2024-08-22)[2024-10-09]. https://www.cnii.com.cn/rmydb/202408/t20240822_595284.html.
[13]	赵海燕, 易庆奥, 汤敬华, 等. 分布式模型训练中的通信优化方法:现状及展望[J/OL]. 小型微型计算机系统, 1-16.(2024-05-30)[2024-10-06]. http://kns.cnki.net/kcms/detail/21.1106.TP.20240529.1438.004.html.
[14]	张宏科, 于成晓, 权伟, 等. 融算网络体系基础研究[J]. 电子学报, 2022, 50(12):2928-2934. doi: 10.12263/DZXB.20221140