A Cloud-based Platform and
Infrastructure Integrating Materials Simulation, Data, HPC, and AI
Xiaoyu Yang1*, Xinjie Ma2, Lifang Xu2, Xiaoneng. Ran2,
Zhongxin Yu2, Zhiyao Bo2, Bo Liu2
1 Computer Network Information Center, Chinese
Academy of Sciences
2 Beijing Maigao Matcloud Technology Co. Ltd
ABSTRACT: The challenge of using
artificial intelligence to help with photoelectron catalysis materials design
is the lack of data. While quantum mechanical simulation can be used to produce
materials data, the difficulties of using it to produce data still barrier many
users. For example, users have to understand Linux, spend time to find
computational resources, and storage resources, etc. In particular, once
simulation completes, the core materials data has to be extracted from the simulation
result. Also, storing the extracted data into the database for sharing and well
management requires extra work.
In order to address this need, we
developed MatCloud. MatCloud is a Cloud-based computational infrastructure for
the integrated management of materials simulation, data and computing
resources. It is directly connected to a computing cluster and a materials
simulation database, integrating the computing facilities, data, various
scripts, and simulation code together to automatically manage the creation and
running of simulation jobs, the subsequent extraction of core output
information, and the longer-term archival of materials properties data. One of
important novelties of MatCloud is that it provides a graphical user interface
for end user to create a customised workflow for running materials simulations.
Once simulation completes, the required materials properties have been acquired
and preserved in the material property database. The more users use MatCloud, the more
simulation data in MatCloud will be accumulated (for using licensed software
such as VASP user must provide license). If users do not wish their data open
to public, they can set their data not searchable by others.
This talk illustrates challenges of
high-throughput simulations, the development of MatCloud, and how MatCloud
supports the integrated management of materials simulation, data and computing
resources.
Keyword: High-throughput Simulation Infrastructure; MatCloud+; Materials Informatics; Integration of Materials simulation, data, HPC and AI.
Prof. Xiaoyu Yang is currently working at Computer Network Information Center, Chinese Academy of Sciences (CAS). He joined CAS in 2012, and awarded the “100 Talent Program” fellowship of Chinese Academy of Sciences. Prof Yang’s research interests currently focus on Material Genome Initiative, which includes: materials informatics, materials genome initiative informatics, materials simulation and data infrastructure etc. Prof. Yang completed his post-doctoral research at the University of Cambridge, UK in 2008. He was previously a Research Associate in the Department of Earth Sciences and affiliated software engineer in Cambridge e-Science Centre, at the University of Cambridge. He joined School of Electronics and Computer Sciences, University of Southampton, UK in 2008, as a Research Engineer. In 2010, he worked in Reading e-Science Center, University of Reading. Prof. Yang has earned an MSc degree in IT (2001) and a PhD degree in Systems Engineering (2006) from "Faculty of Computing Science and Engineering" at the De Montfort University, UK.