1 / 1

An OpenDX-Driven Molecular Dynamics Visualization Solution

OpenDX. MD Code. Bridge Initialization Data. Bridge. Updated Positions. Initial Positions. An OpenDX-Driven Molecular Dynamics Visualization Solution. L.N. Bui, J. Wright, and A.J. Pounds Departments of Chemistry and Computer Science Mercer University. Introduction

Download Presentation

An OpenDX-Driven Molecular Dynamics Visualization Solution

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. OpenDX MD Code Bridge Initialization Data Bridge Updated Positions Initial Positions An OpenDX-Driven Molecular Dynamics Visualization Solution L.N. Bui, J. Wright, and A.J. Pounds Departments of Chemistry and Computer Science Mercer University Introduction The field of molecular dynamics (MD) has been established for years and is used widely in the chemical, physical, medical, and pharmaceutical sciences to study interactions between combinations of atoms and molecules. Depending on the type of potential function used, MD calculations can be extremely time consuming. The largest MD calculations, in fact, are considered “Grand Challenge” problems in computational science circles. The NSF now has a separate category devoted specifically to these types of problems. In traditional MD calculations, the forces on individual atoms and molecules, resulting from the potential set up by the positions of all the other elements in the system, must be explicitly calculated. For each atom or molecule in the system, these forces are integrated to come up with a velocity in 3D space. The motion is then propagated to the next step in time, and the entire process repeated. By far the most time consuming step in this process is the calculation of the forces. This step is O(n2) in computational complexity. There have been numerous techniques used to speed up this process and a current area of interest in Dr. Pounds research is the efficient calculation of MD forces utilizing modern language constructs and parallel computing techniques. In order to verify that the newly designed computational techniques are working properly, it is often necessary to visualize the results of the calculation while it is taking place. This is the impetus for research described in this paper. However, to truly benefit the “science”, we have attempted to create a general tool that can be used by anyone developing MD codes. #include "SHMProxy.h" int main (int argc, char *argv[]) { ... /* extract SHMProxy parameters from command line parameters */ int shmid = strtol (argv[1], NULL, 10); int semid = strtol (argv[2], NULL, 10); int N_ATOMS = strtol (argv[3], NULL, 10); ... SHMProxy shm; // SHMProxy object Molecule m; // Molecule object to hold data ... shm.client_set_shmid (shmid); // set the shared memory ID shm.client_set_semid (semid); // set the semaphore ID shm.client_set_size (N_ATOMS); // set the number of atoms shm.attach (); // attach the shared memory shm.read (m); // read the initial positions while ( /* still running */ ) { ... // do calculations shm.update (m); // write the updated positions ... } ... shm.detach (); // detach the memory before finishing ... } Strategy To allow for the greatest possible audience the software was designed around freely available open source software from the Internet. To visualize the system, the OpenDX (formerly the IBM Data Explorer) data flow visualization system was used along with the Chemistry modules from the Cornell Theory Center. The contribution of the current research is the addition of a “bridge” that connects the users MD code to the visualization system. Again, thinking from the point of view of the MD programmer, the system must be easy to use, only requiring the MD programmer to make a minimum number of calls to the bridge code. Part of the task at hand is building the application program interface (API) for the use of the MD programmer.

More Related