What have we learnt from using real parallel machines to solve real problems?
C3P Proceedings of the third conference on Hypercube concurrent computers and applications - Volume 2
Concurrent multiple target tracking
C3P Proceedings of the third conference on Hypercube concurrent computers and applications - Volume 2
Hi-index | 0.00 |
The concurrent multi-target tracking code is ported to and run on the Intel iPSC and Ncube hypercubes. The (minor) difficulties in this porting are noted. The biggest single problem in running the tracking program is found to be the rather limited node memories on the iPSC and Ncube machines. Timing results for the analysis of a sample threat on the iPSC, Ncube and Mark-III hypercubes are presented. The relative performances of the various machines are found to be consistent with prior expectations. A fairly large sample scenario is investigated using 4 through 8-dimensional Ncube configurations, and variations of Ncube performance for differing levels of CrOS porting are examined. Initial results of running the tracking program on the BBN Butterfly are also presented.