<![CDATA[This research aims to update the full virtualization system (hypervisor) by utilizing GPU passthrough method to enhance system performance. Hypervisor is a software that enables the use of multiple virtual machines on a single physical system. Despite being a popular solution in virtualization environments, challenges related to performance still exist. In this study, the GPU passthrough method is employed to provide direct access to the graphics card for virtual machines, which can significantly improve system performance. The methodology used in this research includes an analysis of virtualization system performance when utilizing GPU passthrough. Several tests and performance measurements are conducted to compare the system's performance with and without the GPU passthrough method. The results of this research are expected to provide a better understanding of the utilization of GPU passthrough method in full virtualization systems and its impact on performance. Furthermore, this study can also offer practical guidance to virtualization system developers in upgrading their systems with more efficient methods. In this context, the GPU passthrough method allows virtual machines to directly utilize graphic resources from the physical graphics card, meaning virtual machines can enjoy higher graphical performance compared to conventional approaches. Using this method provides opportunities to enhance visual quality, speed, and responsiveness of applications run in virtual environments. Overall, this research holds the potential to address performance constraints associated with hypervisors and provide a significant advancement in virtualization technology. By gaining a deeper understanding of the implementation of GPU passthrough, virtualization system developers can optimize their system's performance and deliver a better experience for virtual machine users. It is also hoped that this research will serve as a foundation for further advancements in the field of virtualization and contribute positively to the development of future computing technologies.]]>
Copyrights © 2023
