As science continues to push the envelope of experimentation, computer-based numerical simulation is gaining wide acceptance as a means of quantifying research subjects that are either too large or too small—or move too quickly or too slowly—to be measured by conventional instruments. Indeed, numerical simulation is now not only the domain of mathematicians and theorists, but also of researchers in most life- and physical-science disciplines.
The virtues of numerical simuladon are clear, but problems remain. even] under the most optimal condidons numerical simulation and computer modeling demand large chunks of computer time. The problem is compounded when scientists lack the math or computer science skills to perform numerical simulation efficiently. Traditional mathematics often does not lead to the best computer algorithms, and the average scientist is unable to devote much time to the quest for a better model.
In 1987 the National Science Foundation, recognizing that faster computers alone would ...
