Managing a data structure for allocating graphics processing unit resources to virtual machines

摘要

Methods, techniques, and systems for dynamically allocating graphics processing units among virtual machines are provided. Example embodiments provide a dynamic GPU allocation system (“DGAS”), which enables the efficient allocation of physical GPU resources to one or more virtual machines. In one embodiment, the DGAS comprises a GPU allocation list for use in allocating the physical GPU resources comprising one or more virtual machine entries each containing a designation of a virtual machine, an indication of a GPU benefit factor associated with the designated virtual machine, and an indication of processing bandwidth requirements associated with the designated virtual machine. The entries are ranked based at least upon the GPU benefit factor associated with each designated virtual machine. Available GPU resources are allocated to some subset of these ranked virtual machines as physical GPU capacity is matched with the requirements of the subset.

主权项

1. A computer-implemented method in a server computing system for managing a graphics processing unit (GPU) allocation list for use in allocation of physical GPU resources, including processing bandwidth, to one or more virtual machines, the list comprising one or more virtual machine entries each containing a designation of a virtual machine, an indication of a GPU benefit factor associated with the designated virtual machine, and an indication of processing bandwidth requirements associated with the designated virtual machine, comprising:
rank ordering the one or more virtual machine entries based at least upon the GPU benefit factor associated with each designated virtual machine; determining the processing capacity of one or more physical GPU resources associated with the server computing system; determining a subset of entries on the GPU allocation list that each include a designated virtual machine to which to allocate a portion of the processing bandwidth resources of the one or more physical GPU resources, the determining based upon matching, according to the ranked order of the one or more virtual machine entries, the indicated processing bandwidth requirements associated with each designated virtual machine to the determined processing capacities of the one or more physical GPU resources until no more virtual machines can be accommodated; causing processing bandwidth resources of the one or more physical GPU resources to be allocated to the virtual machines designated by the determined subset of entries on the GPU allocation list; identifying one or more candidate virtual machines designated by entries on the GPU allocation list that are candidates for unseating; determining one or more contender virtual machines designated by entries on the GPU allocation list having corresponding GPU benefit factors that exceed the GPU benefit factors of the candidate virtual machines by some threshold; and de-allocating allocations of processing bandwidth resources from one or more of the candidate virtual machines and allocating the de-allocated processing bandwidth resources instead to the one or more contender virtual machines.