重载图形图像的云原生并行超算技术
Cloud-Native Parallel Supercomputing Technology on the Overloaded Graphics and Images
姚湘军 1易敏 2张焕辉 2范文一 1许家强 1沈方艺 1刘正才3
作者信息
- 1. 湖南天湘和信息科技有限公司,湖南长沙 410000
- 2. 湖南大学国家超级计算长沙中心,湖南长沙,410082
- 3. 湘潭大学土木工程学院,湖南湘潭,411105
- 折叠
摘要
叙述了当代智慧城市建设中3S应用场景、实景三维建模与应用、建筑信息模型和云渲染应用产业等四大领域中的重载图形图像处理任务的现状和所面临的算力、存储和网络等资源需求上的挑战,分析了重载图形图像处理的瓶颈和特点,提出了重载图形图像处理的云原生并行超算技术的基本概念、研究策略和应用解决方案.
Abstract
This paper provides an extensive overview of the current state of overloaded graphics and images processing tasks on four key domains within contemporary smart city de-velopment:3S application scenarios,real-world 3D modeling and applications,building information modeling(BIM),and the cloud rendering application industry. The documents dis-cuss the challenges posed by resource requirements,including computational power,storage,and network,faced by over-loaded graphics and images processing tasks in the context of smart city construction. It analyzes the bottlenecks and charac-teristics specific to overloaded graphics and images processing and introduces the fundamental concepts,research strategies,and application solutions for cloud-native parallel supercom-puting technology on the context of overloaded graphics and images processing.
关键词
重载图形图像/GPU服务器网/云原生/并行超算Key words
overloaded graphics and images/GPU serversnetwork/cloud-native/parallel supercomputing引用本文复制引用
基金项目
国家自然科学基金(62272156)
湖南省湖湘高层次人才聚集工程创新团队项目(2019RS1059)
出版年
2024
NETL
NSTL
万方数据