Space-Time Adaptive Processing on Commercial High-Performance Computers
Abstract
This project consisted of building a portable parallel library for the space-time adaptive processing (STAP) problem. Portability was achieved by using standard like BLAS, LAPACK, SCALAPACK and MPI. The library simplifies implementation of STAP applications on different high-performance parallel computers, allowing rapid prototyping of parallel STAP systems. The library includes common to STAP communication and computation procedure. All library routines take as input 3D data cubes, and produce as outputs also 3D cubes. Two user manuals, one describing the data distributions library, and the other describing the STAP algorithms construction were developed for this project.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 2000
- Accession Number
- ADA382569
Entities
People
- Adam W. Bojanozyk
Organizations
- Cornell University College of Engineering