Graphics Processing Unit Virtualization Enhancement Using Dynamic Partitioning Technique

Abstract

In the era of cloud computing and high-performance computing, performance and efficiency became the core part of the need. As the CPU capabilities are reaching the max GPUs are used to do high-performance computing. GPUs unlocking new possibilities in different sectors including gaming, content creation, deep learning, machine learning, and 3D graphics rendering. People are now shifting from storing everything to their devices and compiling everything in their local device to the central place which can be accessed via the internet and all the processing and storage done by that central place. This holds with gamming also. Due to this server maintainers are facing a lot of issues when it comes to virtualization of the hardware. The report focuses on the need for hardware virtualization and tries to explore the available techniques in the market. The report also proposes a new method that can be incorporated and will enhance the proposed method by using the dynamic partitioning technique.