Xiaoyu Yang (Kevin)
CNIC, Chinese Academy of Sciences
Abstract: For those experiment-centric material scientists who are not familiar with Density Functional Theory (DFT) want to run theoretical simulations, how can they run simulations easily without deep understanding of DFT, or without having to build their computing clusters or buying computing resources? The materials simulation lifecycle data include simulation input and output files, intermediate results files, log and error files, associated metadata, and information capturing the processes that generate the data. How to effectively store and manage them? To address these needs, a high-throughput computational infrastructure for integrated management of material simulation, data and resources, namely, MatCloud, has been developed. In comparison to MaterialStudio and MedeA materials software, MatCloud supports on-line crystal structure modelling, on-line job submission and monitoring through Web browser (for commercial software, user must provide license). In particularly, the simulation output are well preserved and core properties data are extracted saving into the database. Furthermore, MatCloud primarily supports running another quantum-mechanical simulation code ABINIT and micro-scale simulation LAMMPS. MatCloud is a first Cloud-based high-throughput materials simulation and data management infrastructure in China and currently the registered user exceeds 1300. This report illustrates core functionalities, case studies and future vision of MatCloud.
MatCloud高通量材料计算平台与智能化数据管理：
核心功能、案例和展望
杨小渝
中科院计算机网络信息中心
摘要：开展材料计算模拟，不仅要熟悉不同尺度的材料计算软件（如VASP, LAMMPS），还要对Linux系统非常熟悉，及解决计算资源的问题，这极大地增加了开展材料计算模拟的难度，并且计算数据也得不到集中有效地管理，极易丢失，为进一步采用人工智能技术开展材料数据挖掘造成极大不便。
MatCloud是一个将云计算理念（SaaS, PaaS, IaaS）用于材料计算模拟的快捷开展高通量材料计算和形成材料计算数据库的智能化平台，目前注册用户已接近1300。它的开发始于2013年。与业界熟知的Material Studio , MedeA相比，其最大特点之一在于用户无需下载任何软件和插件，只需网页浏览器，通过帐号、密码登陆即可使用（商业软件如VASP，用户需自带版权）。MatCloud可以直接与计算集群或超级计算机相连, 提供图形化建模工具（如掺杂，超胞构建）, 支持高通量筛选（如团簇展开），支持跨尺度的计算流程设计等。实现了大规模材料计算的作业在线提交和监控、结果分析，数据自动处理等功能。此外，MatCloud还初步支持LAMMPS ,ABINIT 等软件，支持声子计算、过渡态搜索等。MatCloud经过多年的迭代开发，已具备较为丰富的功能。 本报告旨在系统整理和剖析MatCloud主要核心功能，高通量筛选案例，及对未来功能扩展的展望。