Computer-Aided Design of Drugs on Emerging Hybrid High Performance Computers

Abstract

The project s overarching objective is to increase the time scale and length scale of scientific simulations relevant to the Army. To this end, the investigator and her group have been working at the University of Delaware (UD) on the algorithmic, implementation, and optimization aspects of large-scale molecular dynamics (MD) simulations on GPUs. During Years 2009-2012, the investigator designed the advance GPUs algorithms for MD simulations and integrated them into an open-source code called FE NZI; during the past year (Year 2012-2013) the investigator has focused her effort on the performance optimization and characterization of the MD code across different GPU generations as well as on the integration of the code into a general framework for non-dedicated, high-end clusters that assures high resource utilization and enables coordinated progressions of MD trajectories.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2013
Accession Number
ADA607433

Entities

People

  • Michela Taufer
  • Narayan Ganesan
  • Sandeep Patel

Organizations

  • University of Delaware

Tags

Communities of Interest

  • Advanced Electronics
  • Biomedical
  • Human Systems

DTIC Thesaurus Topics

  • Algorithms
  • Carbon Nanotubes
  • Chemistry
  • Computational Biology
  • Computer Programming
  • Computer Programs
  • Computer Science
  • Computers
  • Graphics Processing Unit
  • High Performance Computing
  • Linear Algebra
  • Molecular Dynamics
  • Monte Carlo Method
  • Operating Systems
  • Programming Languages
  • Students
  • Systems Biology

Readers

  • Computational Fluid Dynamics (CFD)
  • Parallel and Distributed Computing.
  • Research Science/Academic Research