IMPLEMENTATION OF ASYNCHRONOUS PARALLEL PROCESSING WITH RASPBERRY PI CLUSTER

Abstract

Modern parallel computing has ever been so chunky and expensive. With the use of a miniature

sized Raspberry Pi and Cluster HAT, it becomes more affordable and leaves asmall physical footprint for easy storage. Parallelism can be used to spread the workloadof computation-intensive applications across the multiple cores of a Raspberry Pi. It is the principle upon which modern supercomputers are built. This paper will describe the implementation of asynchronous parallel processing with Raspberry Pi Cluster. This research involves a Raspberry Pi 3, four Raspberry Pi Zeroes, and a ClusterHAT to act as the controller of the cluster. The cluster is then tested against a Python-based modulespecifically used to exploit clusters and a ‘nginx’ web server. The experiment proves thatclustered Raspberries perform better than single ones. It also proves that the cluster seamlessly functions as a single unit regardless of inactive nodes.