Heuristic testing on task partitioning for heterogeneous cluster

Executing large task on a single machine requires CPU to have high processing power. In present environment, most machines are interconnected in a cluster and they are often underutilized. This paper explores heuristically the performance of homogeneous, heterogeneous and multi-core clusters. This work consists of three experiments: Equal task partitioning according to the number of nodes (homogeneous cluster), Equal task partitioning according to the number of nodes (heterogeneous cluster) and Equal task partitioning according to the number of cores (heterogeneous cluster). The task is based on edge detection method (Sobel) which is tested with an array of images. The images are processed in three different sizes; 1K x 1K, 2K × 2K and 3K × 3K. The performance evaluations are based on processing speed as parallel processing measurement. The results yield a list of critical criteria that influence the performance of processing speed in a heterogeneous cluster.