US 12,608,871 B2 · Apparatus and method for efficient graphics virtualization

Abstract

An apparatus and method are described for allocating local memories to virtual machines. For example, one embodiment of an apparatus comprises: a command streamer to queue commands from a plurality of virtual machines (VMs) or applications, the commands to be distributed from the command streamer and executed by graphics processing resources of a graphics processing unit (GPU); a tile cache to store graphics data associated with the plurality of VMs or applications as the commands are executed by the graphics processing resources; and tile cache allocation hardware logic to allocate a first portion of the tile cache to a first VM or application and a second portion of the tile cache to a second VM or application; the tile cache allocation hardware logic to further allocate a first region in system memory to store spill-over data when the first portion of the tile cache and/or the second portion of the file cache becomes full.

Claims33

Dependent claims

A processor comprising:

a graphics processing unit (GPU), the GPU comprising:
a first memory to store data;
a plurality of multiprocessors, each multiprocessor comprising a plurality of execution resources to execute commands in parallel using single-instruction, multiple-thread (SIMT) to process the data, the plurality of execution resources being the same across the plurality of multiprocessors and including arithmetic logic units and load-store units usable to process the data;
a memory interconnect to couple the plurality of multiprocessors to a second memory;
and
allocation hardware logic to allocate a first plurality of the multiprocessors, a first portion of the first memory, and a first portion of the memory interconnect to a first virtual machine (VM), and to further allocate a second plurality of the multiprocessors, a second portion of the first memory, and a second portion of the memory interconnect to a second VM.

3 more elements

A non-transitory machine-readable medium having program code stored thereon which, when executed by a machine, causes the machine to perform:

storing data in a first memory of a graphics processing unit (GPU), the data to be processed by one or more multiprocessors of a plurality of multiprocessors within the GPU, the plurality of multiprocessors coupled to a second memory via a memory interconnect of the GPU, each multiprocessor comprising a plurality of execution resources to execute commands in parallel using single-instruction, multiple-thread (SIMT) to process the data, the plurality of execution resources being the same across the plurality of multiprocessors and including arithmetic logic units and load-store units usable to process the data;
allocating a first plurality of the multiprocessors, a first portion of the first memory, and a first portion of the memory interconnect to a first virtual machine (VM);
and
allocating a second plurality of the multiprocessors, a second portion of the first memory, and a second portion of the memory interconnect to a second VM.

1 more element

Dependencies are inferred from the claim text.

Apparatus and method for efficient graphics virtualization

Life of this patent

Apparatus and method for efficient graphics virtualization

Life of this patent

Abstract

Claims33