题   目：Algebraic multigrid block preconditioners and parallel-in-time algorithms utilizing mixed-precision arithmetic for multi-group radiation diffusion equations

报告人：岳孝强  教授

单   位：湘潭大学

时   间：2024年11月23日10:00

地   点：郑州校区九章学堂南楼C座302

摘   要：Designing robust and accurate numerical solution algorithms in an efficient and scalable manner for multi-group radiation diffusion equations has always been the focus of our intensive research efforts. Algorithms achieving parallelism in time have been of especially high demand as two facts: (i) the higher the spatial resolution signifies a comparative or greater scaling in the temporal direction, (ii) future computing speed must rely on the increased concurrency provided by not faster but more processors. Besides, the server-line products are increasingly featuring low precision special function units, such as the NVIDIA tensor cores in ORNL's Summit supercomputer providing more than an order of magnitude higher performance than what is available in IEEE double precision. In this talk, we propose numerous block preconditioners with algebraic multigrid subsolves and explore the possibility of combining parallel-in-time algorithms and mixed-precision arithmetic to further improve algorithmic concurrency.

简   介：岳孝强，湘潭大学教授、硕博导师。现为FASP、JXPAMG和ParaDiag软件包的核心研发成员，担任美国数学学会《数学评论》评论员，致力于并行代数多重网格法和多重网格时间规约算法的研究工作，曾荣获“高效并行算法与解法器”湖南省优秀研究生导师团队、湖南省计算数学应用软件学会二等青年优秀论文，指导研究生荣获首届“解法器算法及优化竞赛（SolverChallenge21）”特等奖及第二届“解法器算法及优化竞赛（SolverChallenge22）”一等奖、“省优硕士学位论文”和“研究生校长奖优秀奖”等，以第一作者身份在SISC、CiCP、CAMWA、C&F等国内外核心学术期刊上发表论文40余篇，主持了国家科技重大专项XX重大专项项目（课题）、GF基础科研核科学挑战专题、国家自然科学基金青年项目、湖南国家应用数学中心预研项目、面向力学仿真的高性能求解算法研究技术研究项目（华为）、国家野外科学观测站（电科院）及国家能源碳酸盐岩油气重点实验室（中石化）等开放基金。（邀请人：李凌霄）