LL Logo



Space-Time Adaptive
Processing on Commercial

Adam W. Bojanczyk, Gary Adams, Robert Durie, and James Lebak
Cornell University
Department of Electrical Engineering
337 Rhodes Hall
Ithaca, NY 14853-3801
tel: (607) 255-4296
fax: (607) 255-9072
email: adamb@anise.ee.cornell.edu

Abstract We describe a portable library and a programming tool for building high-performance parallel STAP applications. The library contains multiple implementations of functional modules that can be computed using different data distributions. Performance models are computed for each implementation module in the library. The programming tool selects implementations and data distributions for each
functional block in an algorithm to optimize performance for particular dataset sizes and parallel architectures.


LL Logo Disclaimer

Direct comments and questions to: webmaster@ll.mit.edu

MIT Lincoln Laboratory. All rights reserved.