Advanced scientific computer software

With the advent of vector processors over the last ten years, the vector computer has become the most efficient and in some instances the only affordable way to solve certain computational problems. One such computer, the Texas Instruments Advanced Scientific Computer (ASC), has been used extensively at the Naval Research Laboratory to model atmospheric and combustion processes, dynamics of laser implosions, and other plasma physics problems. Furthermore, vectorization is achieved in these programs using standard Fortran. This paper will describe some of the hardware and software differences which distinguish the ASC from the more conventional scalar computer and review some of the fundamental principles behind vector program design. 

Advances in computer science continue to serve as the basis for new extensions to software products. In particular, artificial intelligence techniques have begun to mature to the point at which they can play a role in scientific software. In the future, scientific software will incorporate expert systems technology in order to provide a new level of assistance to scientists in applying statistical and graphical techniques to data analysis. 

The rapid development of quantum chemistry and computer power has led to the birth of a new research area, computational chemistry. This new discipline is just one subfield of the more general area of seientific computing, a relatively new field of research which complements the experimental approach to scientific problems. It should not be confused with the more classical areas of theoretical chemistry and theoretical physics. Experiments are supposed to produce new facts and discoveries theory is supposed to provide a firm general framework for the explanation of observed phenomena through a series of mathematical laws. Scientific computing complements experiment and theory alike it is based on well established theories and formalisms but, like experiments, provides new data. Thanks to the combined use of advanced software and powerful computers, it is now possible to some extent to simulate experiments on a computer, thus often providing a firm basis for the interpretation of experimental results. 

During the last 40 years, computer-based computational methods, such as molecular dynamics and Monte Carlo simulations, have evolved from advanced techniques accessible only to a few speciahsts, to standard calculation approaches available to both scientists and engineers. Similarly, numerical methods of solving the Schrodinger equation, implemented in computational softwares such as Gaussian or Materials studio, are now commonly used to study complex systems. Proper modehng of those systems plays more and more important role in modem scientific research. 

The young scientific discipline Chemometrics has rapidly developed in the past two decades. This enormous increase was initiated by advances in intelligent instruments and laboratory automation as well as by the possibility of using powerful computers and user-friendly software. So, chemometrics became a tool in all parts of quantitative chemistry, but particularly in the field of analytical chemistry. Nowadays, the analyst is increasingly faced with the need to use mathematical and statistical methods in his daily work. 

At the high end of the simulation spectrum considerable human effort is needed and can be motivated by the intrinsic scientific or technical nature of the problem. The software developed should have a long useful lifetime and overall efficiency and speed should be of primary importance [61]. However, the list of very promising parallel computer vendors that are no longer in business has been growing for each year during this decade. Despite this fact, software for programming parallel computers has made a number of important advances in... 

---