Distributed Computation, Numerical Methods, and Scientific Computing for Mathematics and Science Students in an Undergraduate Mathematics Department

点击翻译按钮获取中文摘要

Mathematical Sciences (21) This project tests the educational feasibility of introducing mathematics and science students to mathematical problems whose numerical solution is made possible through the use of distributed computation. Distributed computation, one form of parallel computation, is the process of finding numerical solutions to mathematical problems by distributing the computation over a cluster of processors connected together via an inexpensive network such as a high-speed ethernet network. Recent developments in low-cost computer networking technology have for the first time made the capabilities of distributed computation an economic possibility for mathematics departments at undergraduate institutions. The promise of parallel computing power at an inexpensive price has also led to distributed computation being used increasingly in many fields which have traditionally been clients of the mathematical sciences. The project develops a prototype consisting of educational materials for two newcourses in the current curriculum. The courses introduce students to the fundamentals of distributed computation with as few prerequisites as possible. The first course provides an introduction to numerical analysis and uses the native programming language of Mathematica. The second course introduces students to the mathematical problems that are solvable by means of distributed computation. It uses the native distributed computing capabilities of Mathematica and distributed computing libraries accessible through higher-level programming languages. The project includes development of a computing laboratory to support distributed computation projects.