High Powered and Parallelized Computing with Microcomputers
S. Arora · 2025
In recent years, there has been a growing interest in using microcomputer devices for high-performance computing tasks, particularly in the area of cluster computing. μHPC is a novel paradigm that brings high-performance computing capabilities to resource-constrained environments, leveraging the power of heterogeneous computing platforms and optimizing for low power consumption. By enabling efficient parallel processing of large datasets, μHPC has the potential to revolutionize fields such as bioinformatics, image processing, and edge computing. This analysis focuses on the Raspberry Pi microcomputer specifically and the possibilities surrounding it. However, the use of Raspberry Pi microcomputer clusters for HPC tasks which would be considered truly parallelized is still in its infancy and requires further exploration. One aspect that has received limited attention is the choice of operating system for these clusters. In this paper, we aim to address this gap by performing an analysis of the performance of popular open-source operating systems on Raspberry Pi clusters. We will begin by taking a look at the concept of μHPC in general, but the rest of the paper will also review the history of μHPC, major technologies in the world of μHPC, its improvements overtime, and it's current standing. The goal of this analysis is to determine the best operating system options for highperformance computing on Raspberry Pi clusters, taking into consideration factors such as performance, ease of use, and compatibility with existing HPC software. Obviously, we will focus on lightweight options, the operating systems used for testing in this case will be Raspberry Pi OS, Ubuntu Mate, and Arch Linux. In this case, we do not have access to a Raspberry Pi computer to build a cluster, so this project will be carried out through simulation or emulation environments. For example, cloud-based computing platforms, such as Amazon Web Services (AWS) or Google Cloud Platform (GCP), can be used to set up virtual machines (VMs) that mimic the behavior of Raspberry Pi devices. Another option is to use an emulator, such as QEMU, to run Raspberry Pi images on a different architecture, such as x86, on a personal computer. In either case, the key is to accurately model the behavior of a Raspberry Pi cluster so that the results of the analysis (even if they are just relative) are representative of what would be expected on a real cluster. The results of this project will provide valuable insights for researchers, students, and hobbyists interested in using Raspberry Pi clusters for HPC tasks. The findings will help to better understand the trade-offs and benefits of using Raspberry Pi/similar clusters and opensource operating systems for HPC and provide guidance/predictions for future development in this area.