News Column

Researchers Submit Patent Application, "Virtual Interrupt Delivery from a Graphics Processing Unit (Gpu) of a Computing System without Hardware...

September 4, 2014



Researchers Submit Patent Application, "Virtual Interrupt Delivery from a Graphics Processing Unit (Gpu) of a Computing System without Hardware Support Therefor", for Approval

By a News Reporter-Staff News Editor at Politics & Government Week -- From Washington, D.C., VerticalNews journalists report that a patent application by the inventors Mitra, Surath Raj (Behala, IN); Jia, Neo (Fremont, CA); Wankhede, Kirti (Pune, IN), filed on February 11, 2013, was made available online on August 21, 2014.

The patent's assignee is NVIDIA Corporation.

News editors obtained the following quote from the background information supplied by the inventors: "A computing system (e.g., a server) may include a Graphics Processing Unit (GPU). A hypervisor on the computing system may consolidate Virtual Machines (VMs) on a computing platform of the computing system including the GPU. The VMs may share resources associated with the computing platform. The GPU may be a Peripheral Component Interconnect Express (PCIe)-based device that supports Single Root Input/Output virtualization (SR-IOV). SR-IOV may be designed to deliver interrupts generated in the computing platform to multiple operating system driver stacks. However, the aforementioned delivery may be associated with high implementation costs and/or complexity for devices such as GPUs where high performance is tied to an application state being closely coupled to hardware.

"In the case of a non-SR-IOV based GPU, a single driver stack may execute on the hypervisor, and VMs may be multiplexed on top of the single driver stack. This may allow for interrupt delivery to the hypervisor; however, performance may be reduced because applications executing in the VMs are no longer closely coupled to the GPU hardware.

"Performance may be improved by executing a GPU driver stack in each VM; however, for non SR-IOV based GPU hardware, multiplexing hardware for interrupt delivery may prove to be a challenge. One approach to address the challenge may be to service GPU interrupts in the hypervisor, and, in turn, steer virtual interrupts generated at the hypervisor to the VMs. However, without contextual information normally held within the GPU driver stack, it may be impossible for the hypervisor to actually steer virtual interrupts to the VMs."

As a supplement to the background information on this patent application, VerticalNews correspondents also obtained the inventors' summary information for this patent application: "Disclosed are a method, a device and/or a system of virtual interrupt delivery from a Graphics Processing Unit (GPU) of a computing system without hardware support therefor.

"In one aspect, a method includes loading a driver component on a hypervisor of a computing system including a GPU without hardware support for virtual interrupt delivery. The hypervisor is configured to consolidate a number of Virtual Machines (VMs) on a computing platform of the computing system including the GPU and to virtualize resources thereof. The method also includes loading an instance of the driver component on each of the number of VMs, allocating a memory page associated with work completion by the each of the number of VMs thereto through a driver stack executing on the hypervisor, and sharing the memory page with the driver component on the hypervisor. Further, the method includes delivering, through the hypervisor, an interrupt from the GPU to an appropriate VM based on inspecting the memory page associated with the work completion by the each of the number of VMs.

"In another aspect, a non-transitory medium, readable through a computing system and including instructions embodied therein that are executable through the computing system, is disclosed. The non-transitory medium includes instructions to load a driver component on a hypervisor of the computing system including a GPU without hardware support for virtual interrupt delivery. The hypervisor is configured to consolidate a number of VMs on a computing platform of the computing system including the GPU and to virtualize resources thereof. The non-transitory medium also includes instructions to load an instance of the driver component on each of the number of VMs, instructions to allocate a memory page associated with work completion by the each of the number of VMs thereto through a driver stack executing on the hypervisor, and instructions to share the memory page with the driver component on the hypervisor.

"Further, the non-transitory medium includes instructions to deliver, through the hypervisor, an interrupt from the GPU to an appropriate VM based on inspecting the memory page associated with the work completion by the each of the number of VMs.

"In yet another aspect, a computing system includes a memory, a GPU without hardware support for virtual interrupt delivery, and a hypervisor configured to consolidate a number of VMs on a computing platform of the computing system including the GPU and to virtualize resources thereof. The hypervisor includes a driver component loaded thereon. Each of the number of VMs executes an instance of the driver component loaded thereon. The hypervisor executes a driver stack configured to allocate a memory page associated with work completion by the each of the number of VMs thereto and to share the memory page with the driver component on the hypervisor. The hypervisor is further configured to deliver an interrupt from the GPU to an appropriate VM based on inspecting the memory page associated with the work completion by the each of the number of VMs.

"The methods and systems disclosed herein may be implemented in any means for achieving various aspects, and may be executed in a form of a machine-readable medium embodying a set of instructions that, when executed by a machine, cause the machine to perform any of the operations disclosed herein. Other features will be apparent from the accompanying drawings and from the detailed description that follows.

BRIEF DESCRIPTION OF THE DRAWINGS

"The embodiments of this invention are illustrated by way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements and in which:

"FIG. 1 is a schematic view of a hypervisor-based computing system including a Graphics Processing Unit (GPU) without hardware support for virtual interrupt delivery.

"FIG. 2 is a schematic view of a hypervisor-based computing system configured to handle virtual interrupt delivery from a GPU thereof without hardware support therefor, according to one or more embodiments.

"FIG. 3 is a process flow diagram detailing the operations involved in virtual interrupt delivery from the GPU of the hypervisor-based computing system of FIG. 2, according to one or more embodiments.

"Other features of the present embodiments will be apparent from the accompanying drawings and from the detailed description that follows."

For additional information on this patent application, see: Mitra, Surath Raj; Jia, Neo; Wankhede, Kirti. Virtual Interrupt Delivery from a Graphics Processing Unit (Gpu) of a Computing System without Hardware Support Therefor. Filed February 11, 2013 and posted August 21, 2014. Patent URL: http://appft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&u=%2Fnetahtml%2FPTO%2Fsearch-adv.html&r=179&p=4&f=G&l=50&d=PG01&S1=20140814.PD.&OS=PD/20140814&RS=PD/20140814

Keywords for this news article include: NVIDIA Corporation.

Our reports deliver fact-based news of research and discoveries from around the world. Copyright 2014, NewsRx LLC


For more stories covering the world of technology, please see HispanicBusiness' Tech Channel



Source: Politics & Government Week


Story Tools






HispanicBusiness.com Facebook Linkedin Twitter RSS Feed Email Alerts & Newsletters