Fast Array Processing Of The Laplacian Operator
Original Publication Date: 1980-Feb-01
Included in the Prior Art Database: 2005-Feb-13
Numeric solution of many problems involving differential equations in engineering and physics often uses finite difference techniques. These problems require the evaluation of derivatives of functions over an array of mesh points. Normal relaxation techniques for computing the finite difference approximations to these derivatives operate on each successive array point in a serial manner. These techniques do not take advantage of the parallel processing capability of array processors or languages like APL. A method for parallel processing of the Laplacian is described below.