题目：Research on Digital Twin Technology for Future 6G（面向未来6G的数字孪生技术研究）

作者：L. Li

来源：Intelligent and Converged Networks（智能与融合网络）, vol. 6, no. 1, pp. 20-40.

摘要：With advancement in digitalization and simulation technology, digital twin (DT) technology has emerged as a focal point of research in various industries. In response to the demands for high-quality and efficient wireless communication, it is crucial to conduct an in-depth study of core aspects and key technologies of digital twin technology. This will facilitate a better understanding and exploration of the future development direction of related digital simulation technology within the communication field. This paper provides a systematic summary and analysis of the concept of digital twins, their key technologies, and the current research landscape. Additionally, it explores the research and application fields, as well as the development prospects of digital twin technology in communications. The paper also examines diverse applications of digital twin technology in future 6th generation (6G) networks, including an end-to-end digital twin network architecture framework for non-terrestrial networks (NTNs) in the context of 6G. Finally, it discusses the challenges and opportunities for the widespread implementation of digital twins in future wireless communication networks.

编者译：随着数字化和仿真技术的不断进步，数字孪生技术已成为众多行业研究的焦点。为了满足高质量、高效无线通信的需求，深入研究数字孪生技术的核心环节和关键技术至关重要。这将有助于更好地理解并探索通信领域内相关数字仿真技术的未来发展走向。本文对数字孪生的概念、关键技术以及当前的研究现状进行了系统的总结与分析。此外，还探讨了数字孪生技术在通信领域的研究和应用领域，以及发展前景。本文还考察了数字孪生技术在未来6G网络中的各种应用，包括在6G背景下非地面网络的端到端数字孪生网络架构框架。最后，讨论了数字孪生在未来无线通信网络中广泛应用所面临的挑战与机遇。

题目：公共数据分类分级及应用实践研究

作者：何正庆、吴善鹏、刘超、白惠文、李安伦、吴志刚

来源：大数据,2025,11(03):78-89.

摘要：公共数据分类分级制度是数据基础制度的重要部分，其有力有序的应用实施对数据要素的体系供给、高效流通和安全治理具有重要支撑保障作用。根据江苏省相关部门及设区市分类分级试点情况，提出可行的公共数据分类分级闭环管理方法，并针对不同敏感级别的数据建立分类分级管控体系，在保障数据安全的前提下，优化数据资源配置，促进数据共享开放和授权运营，提高数据流通交易的安全性和透明度，以推动数据要素价值有效释放。

题目：基于超级计算机的高性能计算应用发展现状及趋势研究

作者：刘扬、许建飞、许黄超、吴璨、胡泰源、原惠峰、高凌云、梁文昊、董盛、马英晋、李瑞琳、赵永华

来源：数据与计算发展前沿, 2025, 7(2): 68-85.

摘要：随着信息技术的快速发展和全球数据量的激增，超级计算机（超算）已经成为科学研究和创新发展的重要驱动力。本文旨在探讨超算在多个领域中的应用现状与发展趋势。通过广泛调研全球范围内的超算和领域应用情况，系统性地对相关高性能计算应用进行分类和总结，重点关注化学与材料、物理学等多个领域，探讨相关计算需求与超算的适配和部署情况。此外，本文还积极讨论了网格计算与超算互联。超算在多个领域应用已经展现出了显著的效果。随着应用领域的需要和高性能计算技术的不断发展，对超级计算机的软硬件发展也提出更高要求。虽然超算正处在蓬勃发展的阶段，可应用范围广泛，但本文仅选取了代表性应用领域进行分析总结。超算在加速科学发现和技术创新方面的效率显著提升，为未来的研究和应用提供了强有力的支持。同时，提升超算的性能和适应性将是未来科研进展的重要保障。

题目：Training Large Models on Heterogeneous and Geo-Distributed Resource with Constricted Networks（基于带宽受限网络的大模型在异构且分布式地理资源上的训练研究）

作者：Z. Zong, M. Guo, M. Zhai, Y. Tang, J. Li and J. Zhai

来源：Big Data Mining and Analytics（大数据挖掘与分析）, vol. 8, no. 4, pp. 966-980.

摘要：As the computational demands driven by large model technologies continue to grow rapidly, leveraging GPU hardware to expedite parallel training processes has emerged as a commonly-used strategy. When computational resources within a single cluster are insufficient for large-model training, the hybrid utilization of heterogeneous acceleration hardware has emerged as a promising technical solution. The utilization of heterogeneous acceleration hardware and scheduling of diverse cloud resources have become a focal point of considerable interest. However, these computing resources are often geographically distributed. Due to the lack of awareness of heterogeneous devices and network topologies, existing parallel training frameworks struggle to leverage mixed GPU resources across constrained networks effectively. To boost the computing capability of the connected heterogeneous clusters, we propose HGTrainer, an optimizer designed to plan heterogeneous parallel strategies across distributed clusters for large model training. HGTrainer can adaptively saturate heterogeneous clusters because of the expanded tunable parallelism space for heterogeneous accelerators, with the awareness of relatively lower inter-cluster bandwidth. To achieve this goal, we formulate the model partitioning problem among heterogeneous hardware and introduce a hierarchical searching algorithm to solve the optimization problem. Besides, a mixed-precision pipeline method is used to reduce the cost of inter-cluster communications. We evaluate HGTrainer on heterogeneous connected clusters with popular large language models. The experimental result shows that HGTrainer effectively improves 1.49x training throughput on average for the mixed heterogeneous cluster compared with the state-of-the-art Metis.

编者译：随着大模型计算需求的快速增长，利用GPU硬件加速并行训练过程已成为一种常用的策略。当单个集群内的计算资源不足以满足大模型的训练需求时，异构加速硬件的混合利用作为一种有前景的技术解决方案应运而生。然而，异构计算资源通常在地理位置上分散。由于缺乏对异构设备和网络拓扑结构的认知，现有的并行训练框架在带宽受限的网络环境下难以有效地利用混合GPU资源。为了提升连接的异构集群的计算能力，本文提出了优化策略——HGTrainer，能够自适应地使异构集群饱和。为此，本文对异构硬件之间的模型划分问题进行了建模，并引入了一种分层搜索算法来解决优化问题。同时，采用混合精度流水线方法来降低集群间通信的成本。本文还使用流行的大语言模型对HGTrainer进行了评估。实验结果表明，与最先进的Metis相比，HGTrainer平均有效地提高了1.49倍的训练吞吐量。

题目：面向广域分布式智能计算的运行时算力网络资源协同调度方法研究

作者：宋尧、宋平、高巍、刘述、霍志胜

来源：大数据,2025,11(03):3-16.

摘要：随着人工智能等新一代信息通信技术飞速发展，广域分布式智能计算环境已成为一种重要基础设施。针对广域分布式智能计算环境中资源的高效协同调度难题，提出了一种面向广域分布式智能计算的运行时算力网络资源协同调度方法。该方法设计了关键任务决策与回填、基于关键流量调度的执行保障、数据自适应布局等策略，通过综合分析算力网络中的算、网、存资源使用情况，协同应用3类策略以优化运行时资源的全局利用。实验结果表明，相较于已有方法，该方法可有效提升系统吞吐量，并优化全局数据迁移开销。