Volume 7, Issue 1 -
Spring/Summer 1999

Volume 6, Issue 3
Fall 1998

Volume 6, Issue 2
Spring/Summer 1998

Volume 6, Issue 1
Winter 1998

Volume 5, Issue 4
Fall 1997

Volume 5, Issue 3
Summer 1997

Volume 5, Issue 2
Spring 1997

Volume 5, Issue 1
Winter 1997

Volume 4, Issue 4
Fall 1996

Volume 4, Issue 3
Summer 1996

Volume 4, Issue 2
Spring 1996

Volume 4, Issue 1
Winter 1996

Volume 3, Issue 4
Fall 1995

Volume 3, Issue 3
Summer 1995

Volume 3, Issue 2
Spring 1995

Volume 3, Issue 1
January 1995

Volume 2, Issue 4
October 1994

Volume 2, Issue 3
July 1994

Volume 2, Issue 2
April 1994

Volume 2, Issue 1
January 1994

Volume 1, Issue 4
October 1993

Volume 1, Issue 3
July 1993

Volume 1, Issue 2
April 1993

Volume 1, Issue 1
January 1993

Parallel Molecular Dynamics Simulations

Kim Andrews, Charles Koelbel, Rice University; B. Montgomery Pettitt, University of Houston

Molecular dynamics (MD) is a computationally intensive method of studying the natural time-evolution of a system of atoms using Newton's classical equations of motion (a set of coupled ordinary differential equations). For biologically interesting macromolecular systems, the atoms' motion is governed by interatomic forces that depend on the location of the rest of the atoms in the system. Atomic positions and velocities are repeatedly updated subject to this field of interatomic forces, leading to the time evolution of the (step-wise) particle trajectories. Using this method, the dynamic and thermodynamic behavior of molecular assemblies may be studied, leading toward an understanding of their function and properties at the atomic level.

The wide range of characteristic times for processes that occur in biomolecular systems dictates the length of time that MD simulations must be run to realistically model events of biological interest. While bond stretching only takes 10-20 femtoseconds and determines the time step that may be taken (the stiffness of the ODEs), other processes such as local denaturing may take on the order of seconds. Long, continuous MD simulations are essential when modeling long time-scale phenomena related to catalysis, binding, dielectric effects, diffusion, viscosity, heat capacity, longtime structural reconfiguration as occurs in membranes, proteins, and DNA or other polymers. These simulations are also essential whenever sufficient statistics must be generated to associate microscopic characteristics with time-dependent macroscopic traits of a system. In addition, many interatomic forces act over very long distances, generating more CPU cycles as the number of atoms that influence the motions of a given atom increases.

Indeed, molecular dynamics has always been limited more by the current available computational technology than by investigators' ingenuity. Researchers in the field have typically focused their efforts on simplifying models and identifying what may be neglected while still obtaining acceptable results. This has led to much skepticism on the ability of MD to be used as a predictive tool for experimental work. Parallel computers may hold the key to making biologically relevant calculations tractable without compromise. It is the goal of this project to 1) port and optimize user-friendly MD code to the Intel Paragon, 2) perform realistic calculations in order to validate molecular dynamics methods against experimental data, 3) investigate the validity of some of the more commonly used simplifications, 4) use the knowledge gained from porting the code to assist in compiler design, and 5) extend what the project researchers have learned to other scientific applications.

While there is a plethora of publicly-available molecular modeling programs used to perform MD simulations, the most commonly used programs are CHARMM (Chemistry at HARvard Macromolecular Mechanics), AMBER, and GROMOS. This project is using CHARMM to port to the Paragon because of its versatility, ease of use, and support structure. Currently, the researchers have CHARMM running on the Intel iPSC/860 and the KSR-1 and are performing preliminary scaling tests on the Paragon version. The project's researchers from Rice University include those from both the CRPC and the Keck Center for Computational Biology, a collaboration between Rice and the Baylor College of Medicine.

Table of Contents

News | From the Director | Parallel Profile | Research Focus |
Work in Progress | Education Resources | Resources | Calendar | CRPC Home