PARALLEL-PROCESSING SUPPORT MODULE

BECAUSE SOFTWARE HAS BECOME A DOMINANT FACTOR INSIGNAL-PROCESSING R&D, PRODUCTIVITY DEPENDS HEAVILY ON SOFTWARE ENGINEERING ISSUES: EASE OF WRITING NEW SOFTWARE, RE-USING EXISTING SOFTWARE, AND EXCHANGING TECHNOLOGY BETWEEN DIFFERENT GROUPS. RESEARCHERS WHO USE HIGHLY PARALLEL SIMD MACHINES, OR OTHER MACHINES WITH UNCONVENTIONAL ARCHITECTURES, HAVE DIFFICUTLY IN EXCHANGING TECHNOLOGY WITH USERS OF DIFFERENT MACHINES, EVEN OF THE SAME GENERAL CLASS. TO ADDRESS THIS PROBLEM, A NEW PARALLEL-PROCESSING SUPPORT MODULE WILL BE DEVELOPED TO EXTEND THE COMMERCIALLY SUCCESSFUL ENTROPIC SIGNAL PROCESSING SYSTEM (ESPS). THE DESIGN OF THIS MODULE WILL USE ADVANCED SOFTWARE-ENGINEERING TECHNIQUES TO FACILITATE WRITING SIGNAL-PROCESSING APPLICATIONPROGRAMS SO THAT THEY CAN MAKE EFFECTIVE USE OF A PARALLEL MACHINE AND ARE EASILY PORTABLE FROM ONE MACHINE TO ANOTHER. THESE TECHNIQUES, WHICH INCLUDE ABSTRACT INTERFACES, OBJECT-ORIENTED PROGRAMMING, AND SELF-DESCRIBING OBJECTS, HAVE ALREADY BEEN APPLIED TO SIGNAL PROCESSING IN ESPS AND ITS GRAPHICAL INTERFACE, WAVES+, WHICH ARE RAPIDLY BECOMING POPULAR AT MANY OF THE WORLD'S LEADING SIGNAL-PROCESSING CENTERS. ESPS AND WAVES+ WILL SERVE AS A TECHNOLOGY BASE FOR THE PARALLEL-PROCESSING SUPPORT MODULE. PHASE I WILL INCLUDE A DEFINITION OF REQUIREMENTS AND DESIGN OF THE MODULE INTERFACE. ANTICIPATED BENEFITS/POTENTIALCOMMERCIAL APPLICATIONS - THE PARALLEL-PROCESSING SUPPORT MODULE WILLBENEFIT THE FEDERAL GOVERNMENT BECAUSE OF THE CONSIDERABLE SIGNAL-PROCESSING R&D WORK ITS SUPPORTS, INCLUDING THE COMPUTATIONALLY INTENSIVESORT THAT BENEFITS FROM THE USE OF HIGHLY PARALLEL MACHINES. THE PARALLEL-PROCESSING SUPPORT MODULE WILL ENABLE THIS WORK TO PROCEED FASTER AND MORE EFFICIENTLY. BECAUSE THE PARALLEL-PROCESSING SUPPORT MODULE WILL BE DEVELOPED AS AN EXTENSION TO AN EXISTING COMMERCIAL PRODUCT ALREADY IN WIDE USE IN SIGNAL-PROCESSING LABORATORIES, THE PROBABILITYOF