Graphics Processing Unit Partitioning for Virtualization

A method implemented by a host device having a graphics processing unit (GPU), the method comprising:

• receiving, at the host device, a request for a virtual machine with GPU functionality represented by a model with a plurality of segments that individually represent a respective amount of a GPU capability of a video encoding capability, a video decoding capability, a GPU random access memory (RAM) capability, or a GPU computing capability, the request including data specifying a requested amount of one of the GPU capabilities;
• determining whether the GPU of the host device is available to provide the requested GPU functionality based on the specified amount of the one of the GPU capabilities;
• responsive to a determination that the GPU is available to provide the requested GPU functionality, allocating an available portion of the GPU functionality of the GPU to the virtual machine and configuring the virtual machine with a GPU partition having the requested GPU functionality;
• and
• providing access to the virtual machine with the GPU partition having the requested GPU functionality in response to the received request.

A host device, comprising:

• a graphics processing unit (GPU) to furnish GPU functionality represented by a model with a plurality of segments individually indicative of corresponding GPU capabilities;
• a GPU driver to indicate the GPU functionality provided by the GPU according to the different GPU capabilities;
• a processor;
• and
• a memory having instructions stored thereon that are executable by the processor to cause the host device to:
• determine whether the GPU is capable to handle a request for the GPU functionality by querying the GPU driver regarding whether the GPU has one or more of the GPU capabilities specified in the request, the request specifying an amount of the individual one or more GPU capabilities;
• allocate a portion of the GPU functionality of the GPU to a virtual machine to configure the virtual machine with a GPU partition having the allocated portion of the GPU functionality;
• and
• providing access to the virtual machine having the GPU-partition in response to the request.

A method implemented by a host device hosting a virtual machine and a virtual device, the method comprising:

• receiving, from the virtual machine and at an input of the virtual device, a request requesting GPU functionality furnished by a GPU allocated to the virtual machine, the request indicating an amount of a GPU capability of one of a video encoding capability, a video decoding capability, a GPU random access memory (RAM) capability, or a GPU computing capability;
• determining a data input/output (I/O) hardware work queue of the GPU that is associated with the input of the virtual device based on a mapping of virtual-device inputs to data I/O hardware work queues of the GPU;
• providing GPU-input data included with the request to the GPU via the determined data I/O hardware work queue to enable the GPU to furnish the requested GPU functionality using the provided GPU-input data;
• obtaining, by the virtual device, GPU-processed data from the GPU that results from the GPU functionality furnished by the GPU;
• and
• outputting, by the virtual device, the GPU-processed data to the virtual machine.