算力资源

Search documents
国联股份:集合采购与拼单团购结合优势,新疆算力资源计划今年部署
news flash· 2025-06-25 10:16
Group 1 - The core viewpoint of the article highlights the advantages of combining collective procurement and group buying, which allows for cost reduction through centralized negotiation and scale advantages [1] - The company utilizes a platform that integrates both collective procurement and group buying to achieve low inventory and high turnover by reversing orders [1] - In terms of computing power, the company has registered part of its computing resources in Xinjiang and plans to complete some deployments this year, with future supply aimed at various large models and vertical models [1]
工业和信息化部推动算力互联互通 让算力像水、电一样便捷使用
Ren Min Ri Bao· 2025-06-04 21:40
算力,是推动经济社会高质量发展的新质生产力。日前,工业和信息化部印发《算力互联互通行动计 划》,推动更好整合利用各类公共算力资源,促进算力高效供需匹配,提升我国公共算力使用效率和服 务水平,将有力支撑我国数字经济高质量发展。 将公共算力资源标准化互联,打造应用场景 算力互联互通是在现有互联网体系架构基础上,通过构建统一算力标识、增强异构计算和弹性网络能力 等方式,将不同主体、不同架构的公共算力资源标准化互联,形成可查询、可对话、可调用的服务能 力,实现数据和应用在算力间高效供需匹配、流动互通、迁移计算。 行动计划提出打造算力互联应用场景,包括多个方面,如算力资源服务、任务调度、市场交易、开源社 区运营等新业态场景应用,人工智能、科学计算、智能制造、远程医疗、视联网等企业级场景应用,智 能驾驶、云渲染、云电脑、云游戏等消费级场景应用,以及推动算力互联与能源互联网、工业互联网、 移动互联网等融合创新应用。 促进算力资源使用效率和服务水平的提升 "算力互联互通,将促进算力资源使用效率和服务水平的提升,使算力像水、电一样便捷使用。"工业和 信息化部信息通信管理局有关负责人介绍,人工智能发展催生算力互联需求,用户需要能 ...
我国启动算力互联网试验网建设
news flash· 2025-05-17 06:53
Core Points - The China Academy of Information and Communications Technology (CAICT) has announced the launch of the national computing power internet experimental network during the 2025 World Telecommunication and Information Society Day event [1] - The initiative involves collaboration with major telecom operators including China Telecom, China Mobile, and China Unicom to interconnect their computing resources and facilitate user access to computing power [1] - The computing power internet aims to enhance interconnectivity, tackle key technologies, improve standard systems, and expand industry applications, ultimately increasing the accessibility of computing resources across various sectors [1] Summary by Categories Network Construction - The experimental network will focus on various computing resources such as general computing, intelligent computing, and supercomputing, as well as cloud, edge, and terminal computing [1] - The project has already identified resources from 131 enterprises, aggregating a total of 111.3 billion billion floating-point operations per second in intelligent computing resources [1] User Accessibility - Users will have the flexibility to purchase computing power resources on a pay-per-use basis, similar to how electricity is billed by kilowatt-hours, allowing them to buy exactly what they need [1] - Initial results have been achieved, with trial operations conducted in cities like Beijing and Shanghai, completing nearly 300 task dispatches [1] Standardization and Progress - Breakthroughs have been made in the standardization verification of the computing power identification gateways of the three major telecom operators [1]