Embedded Supercomputing for Edge Devices: A Comprehensive Review

Abstract

Embedded supercomputing has emerged as a promising technology for enhancing the performance and energy efficiency of edge devices such as Internet of Things (IoT) devices, smartphones, and embedded systems. By integrating high-performance computing capabilities in a small form factor, embedded supercomputers can enable real-time data processing, machine learning, and image recognition in edge devices, which can significantly improve their functionality and value. This review provides an in-depth analysis of the current state-of-the-art research and developments in embedded supercomputing for edge devices. It covers various aspects of the field, including hardware architecture, software frameworks, benchmarking, power management, applications, as well as security and privacy. This chapter begins by describing the hardware architecture of embedded supercomputers, including the processing units, memory hierarchy, and interconnects. It then provides an overview of the software frameworks for developing applications on embedded supercomputers, including programming models, software tools, and libraries. This is followed by a discussion of the benchmarking methodologies used for evaluating the performance and energy efficiency of embedded supercomputers, including a description of the power management techniques used for optimizing their energy consumption. The review then provides a detailed analysis of the applications of embedded supercomputing in various fields, such as real-time data processing, machine learning, and image recognition. Finally, it describes the security and privacy challenges of embedded supercomputing and the mechanisms used for addressing them. Overall, it highlights the key advancements, challenges, and future directions in this area, and identifies opportunities for further research and innovation.