Comparison of Conjugate Gradient Method and Jacobi Method Algorithm on MapReduce Framework
As the volume of data continues to grow across many areas of science, parallel computing is a solution to the scaling problem many applications face. The goal of a parallel program is to enable the execution of larger problems and to reduce the execution time compared to sequential programs. Among parallel computing frameworks, MapReduce is a framework that enables parallel processing of data on collections of commodity computing nodes without the need to handle the complexities of implementing a parallel program. This paper presents implementations of the parallel Jacobi and Conjugate Gradient methods using MapReduce. A performance analysis shows that MapReduce can speed up the Jacobi method over sequential processing for dense matrices with dimension ≥ 14,000.
Applied Mathematical Sciences, vol. 8, no. 17, pp. 837-849, 2014
Muhamad Fitra Kacamarga, Bens Pardamean, James Baurley
