GPU resource sharing method and device

Abstract

The invention relates to a GPU (Graphics Processing Unit) resource sharing method and device. The method comprises the following steps: comparing GPU physical resource requirements of a virtual machine with unallocated resources in a resource pool; if the unallocated resources in the resource pool do not meet the GPU physical resource requirements of the virtual machine, adjusting the GPU physicalresources allocated to the virtual machine according to the utilization rate of the GPU physical resources allocated to the virtual machine, and allocating the GPU physical resources to the virtual machine according to the adjusted unallocated resources in the resource pool; otherwise, allocating GPU physical resources to the virtual machine according to the GPU physical resource requirements ofthe virtual machine. According to the GPU resource sharing method and device provided by the invention, a plurality of virtual machines can directly use the GPU resources, the computing power of the virtual machines is improved, and the utilization rate of the GPU resources is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of resource sharing, and in particular relates to a GPU resource sharing method and device. Background technique [0002] With the rapid development of science and technology, in order to meet the growing computing needs, the computing needs of various laboratories are becoming more and more prominent. Ordinary scientific research laboratories are still unable to configure GPU devices for everyone. Therefore, sGPU technology solves the problem of GPU sharing. Use the puzzle. [0003] Currently, technologies for realizing GPU resource sharing include GPU virtualization technology and GPU pass-through technology. GPU pass-through technology can only be bound to one virtual machine for use. A single virtual machine completely monopolizes the GPU device, and other virtual machines cannot use it. A single virtual machine cannot fully use GPU resources, resulting in a great waste of resources. . GPU virtualizat...

AI Technical Summary

Problems solved by technology

GPU pass-through technology can only be bound to one virtual machine for use. A single virtual machine completely monopolizes the GPU device, and other virtual machines cannot use it. A single virtual machine cannot fully use GPU resources, resulting in a great waste of resources.

GPU virtualization technology virtualizes multiple GPU devices and assigns them to multiple virtual machines. GPU resource allocation and occupancy are consistent. Coupled with the limitation of IO virtualization performance problems, it is impossible to achieve reasonable allocation of resources and effective use of resources.

Images