International Journal of Information and Communication Technology Research

International Journal of Information and Communication Technology Research>> Call for Papers(CFP)>>Volume 7, Number 7, July 2017

International Journal of Information and Communication Technology Research


Performance Analysis of High Performance k-Mean Data Mining Algorithm for Multicore Heterogeneous Compute Cluster

Full Text Pdf Pdf
Author(s) Ramesh Singh Yadava, P.K.Mishra
On Pages 390-394
Volume No. 2
Issue No. 4
Issue Date April, 2012
Publishing Date April, 2012
Keywords MPI (Message Passing Interface), MPICH2 (High performance & widely portable implementation of MPI)



Abstract

In this paper, we have study the performance of k-Mean data-mining algorithm (k-Mean),which is implemented on the heterogeneous compute cluster with the multi core programming. The multicore program is implemented with MPI and C for the parallel computing and utilizing the maximum compute power of the heterogeneous cluster. The heterogeneous cluster is established with the help of MPICH2. We have also analyzed the efficiency and performance of k-Mean data mining algorithm for the large dataset. The dataset, which we have used, is chess.txt [1]. The dataset is divided into the number of cores and core compute the dataset independently and makes a data cluster of similar dataset on each processor core. Through this implementation, we have justified that the communication time among the processor cannot be negligible for large dataset. So, the compute time for same dataset on different processors core with same speed and memory is different and the different processor with different speeds and memory access also take different time.

Seperator
    Journal of Information and Communication Technology | Journal of Science and Technology     
Copyrights
2012 IRPN Publishers