Evaluation of the Performances of a Parallel Algorithm to Recognize the Patterns in Relation with the Sequential Variant

Abstract

To achieve an algorithm to efficiently recognize the patterns to return the best solution, we need an intensive processing of the entrance set data. In most cases there is a compromise between the taking over of the entrance set data and the algorithm execution time. The intensive processing of the entrance set data needs rather high calculation resources which cannot always be obtained from the ordinary calculation systems. In this paper we suggest the parallelization of an algorithm to recognize the patterns in scientific literature, its parallelization and the evaluation of this variant in relation to the sequential algorithm. This algorithm will be tested using a cluster formed of 28 nods, each nod having 2 quad core 2.33GHz processors.