Home » Technology » Scientific Computing » The Categorization of Scientific Computing

The Categorization of Scientific Computing

© Adam Hughes

As briefly mentioned in the first article in this series, Scientific Computing emcompasses a wide variety of concepts and means many things to many people. This installment presents a categorization of these topics as one person views them, along with a short discussion about the significance of each and the interplay among all of them. Future articles will most likely deal with issues specific to these categories, though some tweaking may be done along the way.

Computer Simulation

Computer simulation is the foundation of most of the scientific computing practiced in the research community today. Bolstered by rapid advancements in computer technology (as evidenced by both faster performance and cheaper prices), computer simulation techniques have experienced growth which probably could not have been predicted just a decade ago. This situation is quickly transforming simulation from an esoteric pursuit to a viable scientific method. The basic goal of any simulation is to employ a model of some aspect of reality in order to gain a better understanding or to make some predictions. Models often take the form of mathematical equations because math, in conjunction with the laws of physics, biology, etc., gives us a nice, clean way to look at the world. Generally, the more information we can include in a model in terms of numbers and equations, the more accurate that model will be. And therein lies the rub : once we have our "picture" of reality set up, something has to be done with all these numbers and equations to produce a useable result. This something is calculation, of course, which can become quite involved and expensive (in terms of the time it takes to complete). Thus, we are always limited by how many calculations we can perform in a given unit of time. These limits however, are continually being expanded by enhanced computing power and "smarter" computer algorithms.

As researchers develop more sophisticated simulation codes to be run on faster computer platforms, the problem sizes that we can examine is being pushed ever upward. In at least some areas, the day when we can conduct an entire investigation from start to finish, from our workstations, may be closer than many believe.

Data Analysis

Hand-in-hand with simulation itself comes data analysis. We need efficient and accurate means of analyzing all of the data produced via computer simulation. Many innovative techniques have been devised for calculating physical properties of materials, life expectancies of populations, and so-on, from data which is seemingly unmanageable. Of

The copyright of the article The Categorization of Scientific Computing in Scientific Computing is owned by . Permission to republish The Categorization of Scientific Computing in print or online must be granted by the author in writing.