Singular value decomposition on the Connection Machine

Leonardo Rub

Author has no known email address

ABSTRACT

The singular value decomposition of a matrix provides us with valuable information about the structure of the matrix. The matrices arising in geophysical data analysis are very large and consequently the resulting singular value decompositions are computationally expensive to obtain. The following implementation on a Connection Machine shows that with appropriate computer hardware and programming, matrix size will not be as much of an issue.