DocumentCode :
2217565
Title :
STAP performance on a ParagonTM/Touchstone system
Author :
Samson, John R., Jr. ; Grimm, David ; Morrill, Kevin ; Andresen, Ted
Author_Institution :
Spacec & Strategic Syst. Oper., Honeywell Inc., Clearwater, FL, USA
fYear :
1996
fDate :
13-16 May 1996
Firstpage :
315
Lastpage :
320
Abstract :
Honeywell´s Embedded Touchstone and Rugged Touchstone systems are COTS-based, hardware equivalent and software compatible manifestations of the Intel Paragon XP/S Supercomputing System. The ParagonTM/Touchstone system is a robust, massively parallel processing architecture capable of supporting a wide range of implementation paradigms for DoD, DoE, and NASA applications. All applications, but airborne and space-based applications in particular can benefit from the highest possible execution efficiency because execution efficiency cost. This paper describes the performance obtained from the implementation of several STAP (space-time adaptive processing) benchmarks and the Real-Time Multi-Channel Airborne Radar Measurements (RTMCARM) application on ParagonTM/Touchstone processing systems using a high level, data parallel, data flow approach. Demonstrated performance is discussed in terms of end-to-end latency, speedup, high sustained throughput, high sustained implementation efficiency, high compute resource utilization efficiency, scalability, software portability and software productivity improvement using Honeywell´s configurator and data flow shell tools
Keywords :
adaptive signal processing; aerospace computing; airborne radar; aircraft computers; data flow computing; parallel architectures; performance evaluation; pipeline processing; radar computing; radar signal processing; real-time systems; software portability; special purpose computers; Honeywell Embedded Touchstone system; Honeywell Rugged Touchstone system; Intel Paragon XP/S Supercomputing System; Paragon/Touchstone system; RTMCARM; Real-Time Multi-Channel Airborne Radar Measurements; STAP performance; configurator; data flow shell tools; end-to-end latency; high compute resource utilization efficiency; high level data parallel data flow approach; high sustained implementation efficiency; high sustained throughput; performance; robust massively parallel processing architecture; scalability; software portability; software productivity improvement; space-time adaptive processing; speedup; Application software; Computer architecture; Costs; Embedded software; Hardware; NASA; Parallel processing; Robustness; Software performance; Software tools;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Radar Conference, 1996., Proceedings of the 1996 IEEE National
Conference_Location :
Ann Arbor, MI
Print_ISBN :
0-7803-3145-1
Type :
conf
DOI :
10.1109/NRC.1996.510700
Filename :
510700
Link To Document :
بازگشت