题 目：Large Scale material Calculations based on GPU

报告人：汪林望 教授

时  间：2015年9月24日（星期四）下午2:00

地  点：中心教学楼610

ABSTRACT

Large scale (e.g, >10,00 or more atoms) self consistent DFT calculations needs a supercomputer station. It takes a long computer time for general CPU computer station. Therefore the large scale material calculations are limited in many research groups. PWmat is a plane wave pseudopotential code for density functional theory (DFT) material simulations. It is an ab initio code, meaning it only needs atomic positions to predict the material properties. It can perform extremely fast plane wave DFT calculations based on GPU machines and our innovative single precision and double precision mixed algorithm. It deploys the state-of-the-art electronic structure calculation methods with many new features and algorithm innovations. It is a powerful tool for ab initio material science simulations, designed for both theoretical and experimental groups. The PWmat has also been fine tunned with the GPU hardware, especially for our hardware/software integrated solution: Mstation. It can perform DFT self-consistent field (SCF) calculation; band structure calculation; density of state calculation; atomic relaxation; molecular dynamics, and nudged elastic band calculations.

Professor Lin-Wang Wang is a Senior Scientist of Lawrence Berkeley National Laboratory, Berkeley. He was graduated from Physics Department, Shanghai JiaoTong University. As a CUSPEA scholar he went to Cornell University in 1985, in which he got his M.A. and Ph.D. degrees. His research interests are large scale nanosystem electronic structure calculations. He has developed a dozen novel methods in electronic structure calculations, some of which are now widely used. These methods pioneered in O(N) method development and pushed the size limit of plane wave pseudo potential calculation for nanosystems from 100 atoms to 106 atoms. He has published 220 papers, which were cited about 10,000 times. The H-index is 52.

联系方式：新葡萄88805官网办公室 （68913163）

邀 请 人：石丽洁 副教授

网    址：http:/