The Tool For Materials Computation Data Acquisition,Coupling And The Database For Materials Computation Platform

The development of computational materials science has greatly accelerated the progress of new materials’ R&D.However,there are a lot of problems in existing materials computation platforms such as only handling the single computation model,the small computation scale and the low efficiency of computing resource utilization.The materials computation platforms can’t transform the data between different materials computation software in the same scale or different scales,And the procedure of materials computation is also isolated in the process of the new materials’ R&D.The High-throughput Multi-scale Material Simulation and Design Platform(Materials Computation Platform)Project aims to integrate the multi-scale materials computation software,achieve and optimize the automated process in the heterogeneous supercomputing grid,offere an open and shared resource system and database,so as to build a sustainable developing cloud service platform.Based on the materials computation platform project,we present the Python computational materials converter(Pymater)library and the materials database for high-throughput multi-scale materials computation.The main works of this paper as follow:(1)We present the Pymater,a robust Python library for the materials computation data transformation between different scale materials computation software and a data acquisition tool for the materials computation results.The Pymater library aims to provide a tool for the data coupling between different computation software in the automatic process of the multi-scale materials computation,and present a tool for the data acquisition and storage in the automatic process of materials computation.(2)We develop the materials database and the Web service system(hereinafter referred to as database platform system)for the materials database.The database platform system provides users with the query and retrieval function of the materials data,and also provides the external materials computation platform the API for the access of the materials computation data.(3)We design and build the MongoDB distributed database cluster.through the tests,we verify the high availability and the horizontal scalability of the MongoDB distributed database cluster.Finally,In the experiment,we compare the performance of MongoDB,Cassandra and HBase distributed database cluster in the material database usage scenario.The experimental results show that the MongoDB database cluster presents the higher performance of the data access with the material database usage scenario.