DocumentCode :
1628432
Title :
Parallelizing the track-target model for the MIMD machine
Author :
Xiong, W.Z. ; Swietlik, Craig
Author_Institution :
Argonne Nat. Lab., IL, USA
fYear :
1992
Firstpage :
674
Abstract :
The tracking-target model serves as a trajectory generation and scenario definition module which can model ballistic missile threats. The authors document a project to implement the tracking-target model on a parallel processing environment. The tracking-target model is a program based on a DEC VAX/VMS platform in FORTRAN. During this parallelization task, a data dependency analysis was conducted and changes to convert the serial version to the equivalent parallel version were analyzed in two separate areas. First, the physical forces acting on the missile were determined to be independent and were tasked to separate processors for parallel execution. Second, the Runge-Kutta method of solving ordinary differential equations was restructured, so that parallel execution could be achieved. The parallel version has exhibited speedups of up to a factor of 6.3 resulting from a shared memory multiprocessor machine
Keywords :
aerospace computer control; differential equations; iterative methods; military equipment; missiles; parallel processing; tracking systems; DEC VAX/VMS platform; FORTRAN; MIMD machine; Runge-Kutta method; ballistic missile threats; data dependency analysis; ordinary differential equations; parallel model; scenario definition module; shared memory multiprocessor machine; track-target model; trajectory-generation module; Computer applications; Differential equations; Gravity; High performance computing; Laboratories; Missiles; Multiprocessing systems; Optimizing compilers; Power system modeling; Trajectory;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Systems, Man and Cybernetics, 1992., IEEE International Conference on
Conference_Location :
Chicago, IL
Print_ISBN :
0-7803-0720-8
Type :
conf
DOI :
10.1109/ICSMC.1992.271694
Filename :
271694
Link To Document :
بازگشت