Communications of the ACM
Special Feature: Five Questions for Berkeley Lab's CRD Director
Lawrence Berkeley National Laboratory Computational Research Division director David Brown says, "Mathematics is the language of science, and in particular, the language that allows science to be put on computers."
Credit: Lawrence Berkeley National Laboratory
Lawrence Berkeley National Laboratory Computational Research Division director David Brown recently discussed the role of applied mathematics in helping supercomputers solve scientific problems, in honor of the U.S. Department of Energy (DOE) supercomputing month in September.
Brown, who spent 31 years working for DOE national laboratories and trained as an applied mathematician, says, "Mathematics is the language of science, and in particular, the language that allows science to be put on computers."
Brown says improved mathematical models and algorithms have played as large a role as computer performance in boosting the performance of large applications on supercomputers, often surpassing the performance increases attributed to Moore’s Law. "These improvements in performance help scientists make much more efficient use of supercomputers and study problems in greater detail," he says.
For example, a mathematical method called adaptive mesh refinement (AMR) enables scientists to automatically focus a supercomputer on the most relevant parts of the problem, providing what one AMR developer describes as a "numerical microscope." AMR is being applied to issues such as melting ice sheets in Antarctica, supernovas, combustion, and national security.
From Lawrence Berkeley National Laboratory
View Full Article
Abstracts Copyright © 2013 Information Inc., Bethesda, Maryland, USA
No entries found