Large Scale Scientific Computation

Large Scale Scientific Computation is a collection of papers that deals with specialized architectural considerations, efficient use of existing computers, software developments, large scale projects in diverse disciplines, and mathematical approaches to basic algorithmic problems. One paper describes numerical treatment of large highly nonlinear two or three dimensional boundary value problems by quadratic minimization techniques applied in many institutions such as in Laboratoire Central des Ponts et Chaussees, Avions Marcel Dassault et Breguet Aviation. Another paper discusses computer-structured design techniques to improve the reliability, efficiency, and accuracy of future production codes. Computer modelling is a potent tool in numerical weather prediction relying on observation, analysis, initialization, and model development. One paper illustrates a systolic algorithm for matrix triangulation, as well as its uses in the Cholesky decomposition of covariance matrices. Another paper describes the Transient Reactor Analysis Code (TRAC) designed to deal with internal flow problems of nuclear reactors. One paper explains the application of large-scale aerodynamic simulation where the programmer can use finite difference techniques in which a large number of mesh points are strategically and orderly placed in the domain of the flow field. The collection is intended for undergraduates in mathematics, programming, computer science, or engineering courses, and designers or researchers involved in industrial facilities, aeronautics, and nuclear design.