GPU servers for computational biomaterial design, evolutionary GPU code optimization, and on-line simulations
Abstract
We propose to acquire state of the art GPU high performance computing servers. They will be used for multiple computing-intensive re search projects. The GPU servers will support in silico design of biomaterials, which is a newly established research focus of a joi nt theory and experimental collaboration at Arizona State University. The aim of the research effort is to realize complex biomateri als by self-assembly. The servers will be used for state-of-the-art GPU simulations and coarse-grained models to identify designs th at can reliably assemble into target material structures, which will then be characterized experimentally. Further, the computing po wer of the GPU servers will be used to develop a new type of optimization scheme for molecular dynamics simulationcodes, based on a recently developed evolutionary computing optimization method that has been shown to improve performance of certain GPU codes by ov er 400%. Development of this optimization technique for molecular simulations will open new possibilities for simulating larger syst ems and longer timescales, with impacts on all areas which use modeling based on molecular dynamics. Finally, the servers will be us ed to educate researcher in the community of nucleic acid nanotechnology how to computationally design and test nanostructures, and part of the GPU computing power will be integrated into a newly developed online nanostructure simulation webserver for students an d researchers to be trained in using molecular simulation tools.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 07, 2021
- Source ID
- N000142112876
Entities
People
- Petr Šulc
Organizations
- Arizona State University
- Office of Naval Research
- United States Navy