计算机科学
虚拟化
操作系统
完全虚拟化
并行计算
云计算
作者
Yi Gao,Zhanpu Cui,Dong He,Long He
标识
DOI:10.1109/ainit65432.2025.11035659
摘要
This scheme adopts the technical route of para virtualization. Through the collaboration of various modules deployed in virtual machines and the virtual machine monitor (VMM), GPU resources are provided inside the virtual machines. In addition, by designing a GPU resource scheduling algorithm based on dynamic time slices, the waste of GPU resources is reduced. Finally, the usability of the virtualization construction of virtualGraphGPU and the fairness of the GPU scheduling algorithm based on dynamic time slices are verified through tests.
科研通智能强力驱动
Strongly Powered by AbleSci AI