Accelerating scientific discovery through computation and visualization II

Journal of Research of the National Institute of Standards and Technology, May-June, 2002 by James S. Sims, William L. George, Steven G. Satterfield, Howard K. Hung, John G. Hagedorn, Peter M. Ketcham, Terence J. Griffin, Stanley A. Hagstrom, Julien C. Franiatte, Garnett W. Bryant, W. Jaskolski, Nicos S. Martys, Charles E. Bouldin, Vernon Simmons, Oliver P. Nicolas, James A. Warren, Barbara A. am Ende, John E. Koontz, B. James Filla, Vital G. Pourprix, Stefanie R. Copley, Robert B. Bohn, Adele P. Peskin, Yolanda M. Parker, Judith E. Devaney

This is the second in a series of articles describing a wide variety of projects at NIST that synergistically combine physical science and information science. It describes, through examples, how the Scientific Applications and Visualization Group (SAVG) at NIST has utilized high performance parallel computing, visualization, and machine learning to accelerate research. The examples include scientific collaborations in the following areas: (1) High Precision Energies for few electron atomic systems, (2) Flows of suspensions, (3) X-ray absorption, (4) Molecular dynamics of fluids, (5) Nanostructures, (6) Dendritic growth in alloys, (7) Screen saver science, (8) genetic programming.

Key words: discovery science; FEFF; FeffMPI; genetic programming; Hylleraas-Configuration Interaction; immersive environments; Lennard-Jones; nanostructures; screen saver science; parallel computing; QDPD; scientific visualization.

1. Introduction

The process of research may be abstracted into three major components as shown in Fig. 1. Increasingly experiment means computational experiment as computers increase in speed and memory. Parallel computing assists in this by providing access to more processors and more memory. Consequently more complex models that run in feasible times become possible. Laboratory experiments as well are becoming parallel as combinatorial experiments become more common. Both of these lead to large datasets where analysis benefits greatly from visualization.

In this paper we describe research collaborations, between members of the Scientific Applications and Visualization Group (SAVG) and scientists in other labs at NIST, that incorporate parallel computing and visualization as integral parts of the research process. The paper is organized as follows. First, in Sec. 2 we describe our immersive environment. 3D immersion greatly assists in understanding data by literally putting the viewer inside their data. It is used in almost every project we do. In Sec. 3 we describe a computational measurement that is not only the most accurate in the world but also extensible to larger systems. Following, Sec. 4 describes computer simulations of complex fluids like suspensions. We present results for concrete. Parallel computations of the near edge structure of clusters of atoms are given in Sec. 5. Replicated data parallelizations of molecular dynamics simulations of fluids are presented in Sec. 6. We describe parallel algorithms for constructing nanostructures as well as v isualization of these structures in Sec. 7. Section 8 details our simulation of the solidification process of a binary alloy. Screen saver science, Sec. 9, is our name for our distributed computing project, designed to harvest the cycles of computers across NIST when they are in screen saver mode. Lastly, in Sec. 10, we discuss the NIST genetic programming system. Currently it is being used to automatically generate functional forms for measurement errors.

2. Immersive Scientific Visualization

The Immersive Visualization (IV) laboratory plays a key role in accelerating scientific discovery. The NIST scientist views their serial/parallel computation results or experimental data with our IV system. The advanced visualization provided by virtual reality techniques in our IV environment provides greater insight into large, complex data sets by allowing the scientist to interactively explore complex data by literally putting the scientist inside the data.

Fully immersive scientific visualization includes: one or more large rear projection screens to encompass peripheral vision, stereoscopic display for increased depth perception, and head tracking for realistic perspective based on the user's viewing direction. The NIST IV laboratory is configured as an immersive corner with two 2.44 m X 2.44 m (8 ft X 8 ft) screens flush to the floor and oriented 90 degrees to form a corner. The large corner configuration provides a very wide field of peripheral vision. It is also important for the sense of immersion for the screens to be flush with the floor. The system fulfills the other immersive characteristics with stereoscopic displays and head/wand tracking hardware.

Large and complex data sets are becoming more commonplace at NIST, as high performance parallel computing is used to develop high fidelity simulations, and combinatorial experimental techniques are used in the laboratory. IV is significantly different from traditional desktop visualization and significantly more effective at illuminating such data sets (1). One analogy for describing this difference is to consider the difference between viewing the fish in a fish bowl and swimming with the fish at their scale. However, the benefits of IV can only be gained when scientists use it. The key ingredient to making IV accessible to scientists is to provide the ability to simply and quickly move their data into the immersive environment.

The primary software controlling the NIST IV environment is an open source system named DIVERSE (Device Independent Virtual Environments--Reconfigurable, Scalable, Extensible) (2, 3). The DIVERSE API (Application Programming Interface) facilitates the creation of immnersive virtual environments and asynchronous distributed simulations by handling the details necessary to implement the immersive environment. The software runs on a variety of display devices from desktop systems to multi-wall stereographics displays with head tracking.