The innovation engine for new materials

Dan Le

Dan Le

Major: 

Chemical Engineering

University: 

University of California, Santa Barbara

Mentor(s): 

Wangqian Miao

Faculty Sponsor(s): 

Xi Dai

Faculty Sponsor's Department(s): 

Materials

Project Title: 

Use mpi4py to Parallelize Tight Binding Planewave Method for Twisted Bilayer Graphene Systems

Project Description: 

Electronic band structure calculation generally involves rigorous computation on a large list of k points, leading to long process times. Twisted Bilayer Graphene (TBG) systems, typically containing more than 10,000 atoms, make it even harder to perform this kind of calculation efficiently. Our goal is to find a solution to reduce the amount of computation time it takes for running this calculation on such large systems. To do this, we integrate mpi4py, which implements Message Passing Interface (MPI) to python, in our newly developed python package to parallelize the k sampling process of Tight Binding Planewave Method. Essentially, we split the band structure computation for different k points into multiple cores to be computed simultaneously, which significantly speeds up the whole process. Our parallelized code shows robust behavior when we scale up the size of the TBG system as well as the number of k points sampled.