SIMULTANEOUS UTILIZATION OF A FIRST GRAPHICS PROCESSING UNIT (GPU) AND A SECOND GPU OF A COMPUTING PLATFORM THROUGH A VIRTUAL MACHINE (VM) IN A SHARED MODE AND A DEDICATED MODE RESPECTIVELY

摘要

A method includes executing a driver component on a hypervisor of a computing platform including a first graphics processing unit (GPU) and a second GPU, and executing an instance of the driver component in the VM. The method also includes providing support for hardware virtualization of the second GPU in the hypervisor and the instance of the driver component executing in the VM, defining a data path between the VM and the first GPU in a configuration register, and defining a data path between the VM and the second GPU in another configuration register. Further, the method includes providing a capability to the VM to utilize the first GPU in a shared mode with one or more other VM(s) and to simultaneously dedicatedly utilize the second GPU based on reading exposed emulated versions of the configuration register and the another configuration register and the support for the hardware virtualization.

主权项

1. A method comprising:
executing a driver component on a hypervisor of a computing platform comprising a first graphics processing unit (GPU) and a second GPU, the hypervisor being configured to consolidate a VM on the computing platform comprising the first GPU and the second GPU and to virtualize resources thereof; executing an instance of the driver component in the VM; providing support for hardware virtualization of the second GPU in the hypervisor and the instance of the driver component executing in the VM; defining a data path between the VM and the first GPU in a configuration register; defining a data path between the VM and the second GPU in another configuration register; reading, through the instance of the driver component in the VM, an emulated version of the configuration register exposed through a hypervisor component executing on the hypervisor corresponding to the VM and an emulated version of the another configuration register exposed through the hypervisor; and providing a capability to the VM to:
utilize the first GPU in a shared mode with at least one other VM based on the data path definition read through the exposed emulated version of the configuration register, andsimultaneously dedicatedly utilize the second GPU based on the data path definition read through the exposed emulated version of the another configuration register and the support for the hardware virtualization of the second GPU in the instance of the driver component executing in the VM.