二维半导体材料的高通量计算及数据管理平台的设计与实现

Design and implementation of a high throughput computing and data management platform for two dimensional semiconductor materials

材料数据的高通量计算是一项复杂、繁琐的过程.利用开源的电子结构计算代码PWscf作为计算引擎,基于PBS的计算任务管理和调度,本团队开发了二维半导体材料基本特性的高通计算平台.该平台以数据库为核心,通过中间件、程序调度等方法,实现计算任务的自动生成和管理,计算过程的监控以及计算结果的提取、分析和收集入库的功能.详细介绍了系统的实现以及基本功能,并以MX2型二维半导体材料为例,说明了其主要的特性.经过本平台计算,获得了二维半导体材料性能的一些变化规律.本平台为材料性能的高通量计算筛选以及基于大数据的材料构效关系的发现奠定了基础.

The high-throughput calculation of material data is a complex and cumbersome process.Using the open-source elec-tronic structure calculation code PWscf as the calculation engine,based on PBS computing task management and scheduling,our team developed a Qualcomm computing platform for the basic characteristics of two-dimensional semiconductor materials.This platform took the database as its core and achieved the automatic generation and management of computing tasks,monitoring of computing processes,and the extraction,analysis,and collection of computing results through middleware,program scheduling,and other methods.This article provided a detailed introduction to the implementation and basic functions of the system,and took MX2 type two-dimensional semiconductor material as an example to illustrate its main characteristics.After calculation on this platform,some changes in the properties of two-dimensional semiconductor materials were obtained.This platform laid the foundation for high-throughput calculation and screening of material properties,as well as the discovery of material structure-ac-tivity relationships based on big data.