EVALUATION OF MASSIVELY-PARALLEL COMPUTING FOR EXHAUSTIVE AND CLUSTERED MATCHED-FIELD PROCESSING

Many computer algorithms contain an operation that accounts for a subs tantial portion of the total execution cost in a frequently executed l oop. The use of a parallel computer to execute that operation may repr esent an alternative to a sheer increase in processor speed. The signa l processing technique known as matched-field processing (MFP) involve s performing identical and independent operations on a potentially hug e set of vectors. To investigate a massively parallel approach to MFP and clustered nearest neighbors MFP, algorithms were implemented on a DECmpp 12000 massively parallel computer (from Digital Equipment and M asPar Corporation) with 8192 processors. The execution time for the MF P technique on the MasPar machine was compared with that of MFP on a s erial VAX9000-210 equipped with a vector processor. The results showed that the MasPar achieved a speedup factor of at least 17 relative to the VAX9000. The speedup was 3.5 times higher than the ratio of the pe ak ratings of 600 MFLOPS for the MasPar versus 125 MFLOPS for the VAX9 000 with vector processor. The execution speed on the parallel machine represented 64% of its peak rating. This is much better than what is commonly assumed for a parallel machine and was obtained with modest p rogramming effort. An initial implementation of a massively parallel a pproach to clustered MFP on the MasPar showed a further order of magni tude increase in speed, for an overall speedup factor of 35.