An embedded platform-based cluster computer

Abstract

Since tasks are becoming more computationally intensive, the demand for parallel programming or cluster computing increases as it does not reduce the computational time. However, the tools and equipment required for cluster computing are expensive and is not readily accessible to those in the academic environment such as students. With the advent of low-cost embedded platforms like Raspberry Pi and Jetson Nano, it is convenient for students to build a computer cluster based on embedded platforms. This enables them to have hands-on experience and learn the fundamentals of cluster computing. In this paper, a cluster computer is built with Jetson Nanos and the performance of the cluster is analysed when running a computationally intensive application in the CPU as well as in the GPU. The number of nodes in the cluster is varied and the results obtained are presented in this paper.