WO2013109451A1 - Para-virtualized high-performance computing and gdi acceleration - Google Patents

Para-virtualized high-performance computing and gdi acceleration Download PDF

Info

Publication number
WO2013109451A1
WO2013109451A1 PCT/US2013/020916 US2013020916W WO2013109451A1 WO 2013109451 A1 WO2013109451 A1 WO 2013109451A1 US 2013020916 W US2013020916 W US 2013020916W WO 2013109451 A1 WO2013109451 A1 WO 2013109451A1
Authority
WO
WIPO (PCT)
Prior art keywords
gdi
vgpu
compute shader
commands
physical gpu
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/US2013/020916
Other languages
English (en)
French (fr)
Inventor
Meher Prasad MALAKAPALLI
Hao Zhang
Lin Tan
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Microsoft Corp
Original Assignee
Microsoft Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Microsoft Corp filed Critical Microsoft Corp
Priority to JP2014553324A priority Critical patent/JP6140190B2/ja
Priority to KR1020147019876A priority patent/KR102013002B1/ko
Priority to CN201380005836.4A priority patent/CN104067230B/zh
Priority to EP13738213.1A priority patent/EP2805231A4/en
Publication of WO2013109451A1 publication Critical patent/WO2013109451A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T15/00—3D [Three Dimensional] image rendering
    • G06T15/005—General purpose rendering architectures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/4555—Para-virtualisation, i.e. guest operating system has to be modified
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/45579—I/O management, e.g. providing access to device drivers or storage

Definitions

  • Computer systems now commonly perform a host of tasks (e.g., word processing, scheduling, accounting, etc.) that prior to the advent of the computer system were performed manually. More recently, computer systems have been coupled to one another and to other electronic devices to form both wired and wireless computer networks over which the computer systems and other electronic devices can transfer electronic data. Accordingly, the performance of many computing tasks is distributed across a number of different computer systems and/or a number of different computing environments.
  • tasks e.g., word processing, scheduling, accounting, etc.
  • Some computer systems are configured to provide para-virtualized execution environments, which allow guest software to share hardware devices of a single computer system in an isolated manner.
  • para-virtualized execution environments provide a plurality of partitions, supported by a hypervisor.
  • the partitions provide isolation between different guest software.
  • the partitions generally include a root partition and one or more child partitions.
  • the root partition runs a host operating system and manages a virtualization stack.
  • the root partition may gain access to physical devices.
  • Each child partition hosts guest software (e.g., guest operating systems and guest applications). Child partitions are provided access to physical devices through virtual devices and software interfaces of the hypervisor.
  • Some para-virtualized execution environments provide child partitions (and guest software executing therein) with para-virtualized access to one or more physical graphics processing units ("GPUs").
  • GPUs physical graphics processing units
  • each implementation of para-virtualized access to physical GPUs supports particular three-dimensional rendering framework(s).
  • guest software may be unable to access capabilities of a physical GPU if that guest software is executing within a para-virtualized execution environment that does not support those capabilities.
  • the guest software may rely on using a virtualized CPU to perform tasks not supported para-virtualized access to a physical GPU, incurring a potentially significant performance penalty.
  • the present invention extends to methods, systems, and computer program products for providing high-performance computing and graphics device interface
  • GDI GDI acceleration in a para-virtualized environment
  • Some embodiments include a method for providing graphics processing unit (“GPU") accelerated computing functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • a virtual machine session is instantiated.
  • a hypervisor in the virtual machine session provides (i) a root partition (which has access to a physical GPU), and (ii) the child partition (which executes the guest application).
  • a virtualized graphics processing unit (“vGPU”) executing within the child partition is presented to the guest application.
  • a user-mode driver (“UMD) of the vGPU presents compute shader device driver interfaces ("DDIs") to the guest application.
  • the compute shader DDIs provide an application programming interface (“API”) that enables the guest application to send compute shader commands to the vGPU.
  • the compute shader commands are used to perform general-purpose graphics processing unit
  • GPU graphics processing unit
  • a render component executing within the root partition receives a physical GPU-specific compute shader command from the vGPU, and schedules the command for execution at the physical GPU.
  • Additional embodiments include a method for providing GPU-accelerated GDI functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • a virtual machine session is instantiated.
  • a hypervisor in the virtual machine session provides (i) a root partition having access to a physical GPU, and (ii) the child partition which executes the guest application.
  • a vGPU which executes within the child partition, is presented to the guest application.
  • An API of a kernel-mode driver (“KMD") of the vGPU enables the guest operating system to accelerate GDI rendering commands submitted by a guest application. These commands are then processed by the KMD of the vGPU.
  • KMD kernel-mode driver
  • a render component executing within the root partition receives a GDI acceleration rendering command from the vGPU.
  • the render component schedules the GDI acceleration rendering command on a GDI composition device within the root partition.
  • the GDI composition device is configured to execute the GDI acceleration rendering command at the physical GPU.
  • the GDI composition device marks a GDI surface corresponding for the GDI command as sharable for composition by the desktop.
  • Figure 1 A illustrates an example computer architecture that enables para- virtualized access to compute shader functionality of physical graphics processing unit (“GPU”) hardware.
  • GPU graphics processing unit
  • Figure IB illustrates an example computer architecture that enables para- virtualized graphics device interface (“GDI”) acceleration by physical GPU hardware.
  • GDI graphics device interface
  • Figure 1C illustrates an example computer architecture that enables para- virtualized access to compute shader functionality and para-virtualized GDI acceleration by physical GPU hardware.
  • Figure 2A illustrates a flow chart of an example method for providing GPU- accelerated computing functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • Figure 2B illustrates a flow chart of an example method for providing GPU- accelerated GDI functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • the present invention extends to methods, systems, and computer program products for providing high-performance computing and graphics device interface
  • GDI GDI acceleration in a para-virtualized environment
  • Some embodiments include a method for providing graphics processing unit (“GPU") accelerated computing functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • a virtual machine session is instantiated.
  • a hypervisor in the virtual machine session provides (i) a root partition (which has access to a physical GPU), and (ii) the child partition (which executes the guest application).
  • a virtualized graphics processing unit (“vGPU”) executing within the child partition is presented to the guest application.
  • a user-mode driver (“UMD) of the vGPU presents compute shader device driver interfaces ("DDIs") to the guest application.
  • the compute shader DDIs provide an application programming interface (“API”) that enables the guest application to send compute shader commands to the vGPU.
  • the compute shader commands are used to perform general-purpose graphics processing unit
  • GPU graphics processing unit
  • a render component executing within the root partition receives a physical GPU-specific compute shader command from the vGPU, and schedules the command for execution at the physical GPU.
  • Additional embodiments include a method for providing GPU-accelerated GDI functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • a virtual machine session is instantiated.
  • a hypervisor in the virtual machine session provides (i) a root partition having access to a physical GPU, and (ii) the child partition which executes the guest application.
  • a vGPU which executes within the child partition, is presented to the guest application.
  • An API of a kernel-mode driver (“KMD") of the vGPU enables the guest operating system to accelerate GDI rendering commands submitted by a guest application. These commands are then processed by the KMD of the vGPU.
  • KMD kernel-mode driver
  • a render component executing within the root partition receives a GDI acceleration rendering command from the vGPU.
  • the render component schedules the GDI acceleration rendering command on a GDI composition device within the root partition.
  • the GDI composition device is configured to execute the GDI acceleration rendering command at the physical GPU.
  • the GDI composition device marks a GDI surface corresponding for the GDI command as sharable for composition by the desktop.
  • Embodiments of the present invention may comprise or utilize a special purpose or general-purpose computer including computer hardware, such as, for example, one or more processors and system memory, as discussed in greater detail below.
  • Embodiments within the scope of the present invention also include physical and other computer- readable media for carrying or storing computer-executable instructions and/or data structures.
  • Such computer-readable media can be any available media that can be accessed by a general purpose or special purpose computer system.
  • Computer-readable media that store computer-executable instructions are computer storage media (devices).
  • Computer-readable media that carry computer-executable instructions are transmission media.
  • embodiments of the invention can comprise at least two distinctly different kinds of computer-readable media: computer storage media (devices) and transmission media.
  • Computer storage media includes RAM, ROM, EEPROM, CD-ROM, solid state drives (“SSDs”) (e.g., based on RAM), Flash memory, phase-change memory (“PCM”), other types of memory, other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer.
  • SSDs solid state drives
  • PCM phase-change memory
  • other types of memory other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer.
  • a "network” is defined as one or more data links that enable the transport of electronic data between computer systems and/or modules and/or other electronic devices.
  • a network or another communications connection either hardwired, wireless, or a combination of hardwired or wireless
  • the computer properly views the connection as a transmission medium.
  • Transmissions media can include a network and/or data links which can be used to carry desired program code means in the form of computer-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer. Combinations of the above should also be included within the scope of computer-readable media. [0027] Further, upon reaching various computer system components, program code means in the form of computer-executable instructions or data structures can be transferred automatically from transmission media to computer storage media (devices) (or vice versa).
  • NIC network interface module
  • computer storage media can be included in computer system components that also (or even primarily) utilize transmission media.
  • Computer-executable instructions comprise, for example, instructions and data which, when executed at a processor, cause a general purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions.
  • the computer executable instructions may be, for example, binaries, intermediate format instructions such as assembly language, or even source code.
  • the invention may be practiced in network computing environments with many types of computer system configurations, including, personal computers, desktop computers, laptop computers, message processors, hand-held devices, multi-processor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile telephones, PDAs, tablets, pagers, routers, switches, and the like.
  • the invention may also be practiced in distributed system environments where local and remote computer systems, which are linked (either by hardwired data links, wireless data links, or by a combination of hardwired and wireless data links) through a network, both perform tasks.
  • program modules may be located in both local and remote memory storage devices.
  • GPGPU Para-Virtualized Compute Shader
  • Figure 1A illustrates an example computer architecture 100a that enables para- virtualized access to compute shader functionality of physical GPU hardware.
  • computer architecture 100a includes physical hardware 101.
  • Physical hardware 101 can include any appropriate hardware devices, such as one or more general purpose processors, system memory, and the like.
  • physical hardware includes at least one physical GPU 101a.
  • Physical GPU 101a is a processing device configured to perform parallel processing tasks, such as graphics rendering.
  • Physical GPU 101a includes support for executing a compute shader, which enables physical GPU 101a to perform general- purpose (i.e., non-graphics rendering) calculations.
  • physical GPU 101a supports GPGPU computation on a compute shader device.
  • Computer architecture 100a also includes hypervisor 102.
  • Hypervisor 102 executes on top of physical hardware 101 and supports a virtualization platform.
  • the virtualization platform provides a plurality of partitions. Each partition provides a logical unit of isolation, in which guest software can be executed.
  • computer architecture 100a includes root partition 103 and child partition 104.
  • Root partition 103 executes a host operating system, and has direct access to physical hardware 101 (as depicted by root partition 103 appearing over physical hardware 101).
  • Each child partition provides an execution environment for executing guest software (e.g., operating systems and/or applications) and may access physical hardware 101 indirectly in a para-virtualized manner. That is, through hypervisor 102, each child partition provides one or more software interfaces (e.g., virtualized hardware) to guest software. The guest software, in turn, uses the software interface(s) to access physical hardware 101.
  • Hypervisor 102 can provide support for a plurality of child partitions.
  • guest software 105 executes within child partition 104.
  • Guest software 105 comprises any appropriate guest software, such as an operating system and/or an application program executing within an operating system.
  • Guest software 105 includes or uses graphics runtime 105 a.
  • Graphics runtime 105 a provides a framework (e.g., APIs) for rendering graphics and/or performing GPGPU computation.
  • Child partition 104 provides guest software 105 access to vGPU 106.
  • vGPU 105 virtualizes physical GPU 101a, enabling guest software 105 to indirectly access physical GPU 101a.
  • vGPU 106 is configured to expose all, or a subset, of the
  • vGPU 106 is configured to expose one or more software interfaces that enable guest software 105 to call vGPU 106 to access compute shader functionality of physical GPU 101a for performing GPGPU computation at physical GPU 101a.
  • vGPU 106 works in conjunction with a render component in root partition 103 to perform compute shader functionality on physical GPU 101a.
  • root partition 103 includes render component 112.
  • Render component 112 in turn, includes compute shader component 113 for handing compute shader commands and data.
  • vGPU 106 remotes compute shader commands and data 117 to render component 112 to perform the rendering on physical GPU 101a.
  • Render component 112 schedules any graphics commands received from vGPU 106 for execution on physical GPU 101a. Render component 112 also creates proper context for executing those commands. As such, render component 112 is configured to use compute shader component 113 to schedule execution of compute shader-related commands that are received from vGPU 106 in child partition 104 on physical GPU 101a.
  • vGPU 106 includes user-mode driver 106a executing in a user- mode of child partition 104 and kernel-mode driver 109 executing in a kernel-mode of child partition 104.
  • User-mode driver 106a exposes device driver interfaces ("DDIs") of at least one rendering framework, including DDIs related to compute shader (GPGPU) functionality, depicted as compute shader DDIs 107.
  • Compute shader DDIs 107 enable guest software 105 to make calls to vGPU 106 for performance of GPGPU computations.
  • user-mode driver 106b exposes DDIs of a rendering framework that supports compute shader functionality (e.g., DirectX® versions 10 and/or 11 from Microsoft® Corporation).
  • DDIs of a rendering framework that supports compute shader functionality
  • user-mode driver 106a may expose the one or more of the following DDIs as part of compute shader DDIs 107:
  • user-mode driver 106b presents all DDIs of a rendering framework that supports compute shader functionality (e.g., DirectX® version 11), such as DDIs related to one or more of a domain shader, a hull shader, or a geometry shader.
  • compute shader functionality e.g., DirectX® version 11
  • user-mode driver 106b may expose any set of DDIs from any rendering framework supporting compute shader functionality.
  • vGPU 106 may also include legacy user-mode driver 106b executing in user-mode of child partition 104.
  • Legacy user-mode driver 106b may expose DDIs of a legacy version of one or more rendering frameworks.
  • legacy user-mode driver 106b may support a legacy version of DirectX® (e.g., DirectX® version 9), or a legacy version any other rendering framework (e.g., OpenGL® from Silicon Graphics, Inc.).
  • user-mode driver 106a is configured to construct hardware contexts and command buffers.
  • user-mode driver 106a converts graphic commands issued by guest software 105 (or graphics runtime 105a of guest software 105) into hardware-specific commands.
  • user-mode driver 106a may receive graphics commands 115 relating to GPGPU computations using a compute shader from guest software 105.
  • User-mode driver 106a is configured to convert graphics commands 115 into hardware-specific commands (i.e., commands that are specific to physical GPU 101a).
  • user-mode driver 106a maintains proper hardware context for physical GPU 101a.
  • user-mode driver 106a translates logical values for settings affecting a graphics pipeline into values and corresponding physical settings.
  • User-mode driver 106a is also configured to store converted hardware-specific commands in command buffer 116 and send command buffer 115 to kernel-mode driver 109.
  • vGPU 106 includes kernel-mode driver 109 executing in kernel- mode of child partition 104, which includes compute shader component 110.
  • Kernel- mode driver 109 is configured to receive command buffers (e.g., command buffer 116) and to construct corresponding direct memory access ("DMA") buffers.
  • command buffers e.g., command buffer 116
  • DMA direct memory access
  • a GPU scheduler calls kernel-mode driver 109. Kernel- mode driver 109 then handles the specifics of actually submitting the DMA buffer to physical GPU 101a.
  • Figure 2A illustrates a flow chart of an example method 200a for providing GPU-accelerated computing functionality to a guest application executing in a child partition of a para-virtualized execution environment.
  • Method 200a will be described with respect to the components and data of computer architecture 100a.
  • Method 200a includes an act of instantiating a virtual machine session, including instantiating a hypervisor that provides (i) a root partition having access to the physical GPU, and (ii) the child partition which executes the guest application (act 201).
  • computing environment 100a can instantiate hypervisor 102 as part of instantiating a virtual machine session.
  • Hypervisor 102 provides root partition 103 and child partition 104.
  • Root partition 103 has access to physical GPU 101a.
  • Child partition 104 executes one or more guest applications, including guest application 105, and has indirect access to physical GPU 101a.
  • Method 200a also includes an act of presenting a vGPU to the guest application, the vGPU executing within the child partition, including presenting a plurality of compute shader DDIs to the guest application as part of a UMD of the vGPU, the plurality of compute shader DDIs providing an API that enables the guest application to send compute shader commands to the vGPU for performing GPGPU computations at the physical GPU using a compute shader (act 202).
  • child partition 104 can present vGPU 106 to guest software 105.
  • vGPU 106 includes user-mode driver 106a.
  • User-mode driver 106a presents compute shader DDIs 107, which enable guest software 105 to call vGPU 106 to perform compute shader GPGPU calculations at physical GPU 101a.
  • guest software 105 or graphics runtime 105 a
  • User-mode driver 106a converts received graphics commands 115 to physical- hardware specific commands (e.g., as part of command buffer 116). vGPU 106 then remotes compute shader commands and data 117 (including physical-hardware specific commands) to render component 112.
  • Method 200a also includes an act of a render component executing within the root partition receiving a physical GPU-specific compute shader command from the vGPU (act 203).
  • render component 112 which executes in root partition 103, can receive compute shader commands and data 117, including a physical GPU-specific compute shader command, from vGPU 106.
  • Method 200a also includes an act of the render component scheduling the physical GPU-specific compute shader command for execution at the physical GPU (act 204).
  • compute shader component 113 can schedule the received physical GPU-specific compute shader command for execution on a compute shader device at physical GPU 101a. In doing so, compute shader component 113 can configure and maintain appropriate context for execution of the physical GPU-specific compute shader command.
  • FIG. 1B illustrates an alternate computer architecture 100b that enables para-virtualized GDI acceleration by physical GPU hardware.
  • computer architecture 100b enables guest software to request GDI command acceleration using a physical GPU, as opposed to handling GDI commands with a physical or virtual central processing unit.
  • child partition 104' computer architecture 100b provides components that enable accelerated rendering of GDI commands on physical GPU 101a' when those commands are issued by guest software 105'.
  • child partition 104' includes GDI interface 108, which is configured to communicate with graphics runtime 105a' (e.g., a graphics runtime that is part of a guest operating system).
  • vGPU 106' is configured to process GDI acceleration commands 120 received from GDI interface 108 at GDI component 11 1 of kernel-mode driver 109'.
  • GDI interface 108 is configured to expose one or more GDI command interfaces to graphics runtime 105a', and to forward GDI accelerated commands received from graphics runtime 105a' to kernel-mode driver 109' of vGPU'.
  • graphics runtime 105a' is enabled to send GDI accelerated graphics commands 120 to kernel-mode driver 109' of vGPU 106' via GDI interface 108 to accelerate GDI command execution at physical GPU 101a'.
  • GDI component 111 of kernel-mode driver 109' is configured to implement one or more GDI command interfaces, along with corresponding hardware rendering operation(s).
  • vGPU 106' implements and exposes GDI commands corresponding to a particular version of DirectX® (e.g., DirectX® versions 10 and/or 11).
  • DirectX® e.g., DirectX® versions 10 and/or 11
  • GDI component 111 may expose and implement a 'DxgkDdiRenderKm' interface, along with corresponding hardware rendering operations.
  • GDI component 111 may implement any appropriate GDI operations, in some embodiments GDI component 111 implements the following operations:
  • Kernel-mode driver 109' is configured to send GDI commands and data 118 to render component 112' within root partition 103'.
  • GDI commands and data 118 include information relating to 3D rendering devices and contexts for executing GDI acceleration commands.
  • Render component 112' includes GDI component 114, which is configured to receive GDI commands and data 118 and to execute received GDI acceleration commands at physical GPU 101a'.
  • GDI component 114 is configured to create a GDI surface and a corresponding 3D rendering (or composition) device (e.g., a D3D device).
  • the 3D rendering device is configured to provide appropriate context for executing GDI accelerated commands.
  • the GDI surface can comprise any appropriate GDI surface, such as any of the following GDI surface types:
  • GDI component 114 marks any created GDI surface as a sharable surface. This enables composition to a desktop to succeed.
  • Figure 2B illustrates a flow chart of an example method 200b for providing GPU-accelerated GDI functionality to a guest application executing in a child partition of a para-virtualized execution environment. Method 200b will be described with respect to the components and data of computer architecture 100b.
  • Method 200b includes an act of instantiating a virtual machine session, including instantiating a hypervisor that provides (i) a root partition having access to the physical GPU, and (ii) the child partition which executes the guest application (act 205).
  • computing environment 100b can instantiate hypervisor 102' as part of instantiating a virtual machine session.
  • Hypervisor 102' provides root partition 103' and child partition 104'.
  • Root partition 103' has access to physical GPU 101a'.
  • Child partition 104' executes one or more guest applications, including guest application 105', and has indirect access to physical GPU 101a'.
  • Method 200b also includes an act of presenting a vGPU to the guest application, the vGPU executing within the child partition, including presenting an API of a KMD of the vGPU that enables a guest operating system to accelerate GDI rendering commands used by the guest application to the vGPU for processing by the KMD of the vGPU (act 206).
  • child partition 104' presents vGPU 106' and GDI interface 108 to guest software 105'.
  • GDI interface 108 enables graphics runtime 105a' (e.g., a graphics runtime of an operating system) to accelerate GDI rendering commands used by guest software 105' to vGPU 106' for processing by kernel-mode driver 109' and for acceleration by physical GPU 101a'.
  • graphics runtime 105a' can send graphics commands 120 to kernel-mode driver 109' through GDI interface 108 for performance of GDI acceleration commands.
  • vGPU 106' uses GDI component 111 of kernel-mode driver 109' to process received GDI commands and to send GDI commands and data 118 to render component 112' in root partition 103'.
  • Method 200b also includes an act of a render component executing within the root partition receiving a GDI acceleration rendering command from the vGPU (act 207).
  • render component 112' which executes in root partition 103', can receive GDI commands and data 118 from vGPU 106'.
  • Method 200b also includes an act of the render component scheduling the GDI acceleration rendering command on a GDI composition device within the root partition, the GDI composition device being configured to execute the at least one GDI acceleration rendering command at the physical GPU, the GDI composition device also being configured to mark a GDI surface corresponding to the at least one GDI acceleration rendering command as sharable for composition by a desktop (act 208).
  • GDI component 114 can create a 3D rendering (composition) device and a GDI surface for execution of GDI acceleration commands and can schedule the GDI acceleration commands for execution on the 3D rendering (composition) device.
  • the GDI surface can be marked as sharable for composition at the desktop.
  • a single computer architecture can provide both compute shader (GPGPU) functionality and para-virtualized GDI acceleration.
  • Figure 1C illustrates an example computer architecture 100c that enables para-virtualized access to compute shader functionality and para-virtualized GDI acceleration by physical GPU hardware.
  • computer architecture 100c includes GDI interface 108 and vGPU 106".
  • vGPU 106 "contains user-mode driver 106a", which includes compute shader DDIs 107 .
  • vGPU 106" also contains kernel-mode driver 109" that includes both compute shader component 110a and GDI component 111.
  • vGPU 106" (executing in child partition 104"), can communicate both compute shader commands and data and GDI commands and data to render component 112" executing in root partition 103".
  • Render component 112" includes both compute shader component 113 and GDI component 114 for executing compute shader and GDI commands at GPU 101a".
  • computer architecture 100c by including both compute shader and GDI functionality, can provide increased functionality and greater compatibility with particular rendering frameworks.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Graphics (AREA)
  • Image Generation (AREA)
  • Processing Or Creating Images (AREA)
  • Stored Programmes (AREA)
PCT/US2013/020916 2012-01-17 2013-01-10 Para-virtualized high-performance computing and gdi acceleration Ceased WO2013109451A1 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP2014553324A JP6140190B2 (ja) 2012-01-17 2013-01-10 準仮想化された高性能コンピューティングおよびｇｄｉ高速化
KR1020147019876A KR102013002B1 (ko) 2012-01-17 2013-01-10 반가상화된 고성능 컴퓨팅 및 gdi 가속 기법
CN201380005836.4A CN104067230B (zh) 2012-01-17 2013-01-10 用于准虚拟化高性能计算和gdi加速的方法和系统
EP13738213.1A EP2805231A4 (en) 2012-01-17 2013-01-10 COMPUTERIZED COMPUTERIZED COMPUTER HIGH PERFORMANCE PARA-VIRTUALIZED AND GDI INTERFACE ACCELERATION

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US13/352,121 US8941670B2 (en) 2012-01-17 2012-01-17 Para-virtualized high-performance computing and GDI acceleration
US13/352,121 2012-01-17

Publications (1)

Publication Number Publication Date
WO2013109451A1 true WO2013109451A1 (en) 2013-07-25

Family

ID=48779647

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2013/020916 Ceased WO2013109451A1 (en) 2012-01-17 2013-01-10 Para-virtualized high-performance computing and gdi acceleration

Country Status (6)

Country Link
US (1) US8941670B2 (enExample)
EP (1) EP2805231A4 (enExample)
JP (1) JP6140190B2 (enExample)
KR (1) KR102013002B1 (enExample)
CN (1) CN104067230B (enExample)
WO (1) WO2013109451A1 (enExample)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2018512661A (ja) * 2015-03-23 2018-05-17 インテル コーポレイション グラフィックス・プロセッサ仮想化のためのシャドー・コマンド・リング

Families Citing this family (25)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102890643B (zh) * 2012-07-26 2015-01-21 上海交通大学 基于应用效果即时反馈的显卡虚拟化下的资源调度系统
US9098323B2 (en) * 2013-09-05 2015-08-04 Nvidia Corporation 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
US9851998B2 (en) 2014-07-30 2017-12-26 Microsoft Technology Licensing, Llc Hypervisor-hosted virtual machine forensics
US9678773B1 (en) 2014-09-30 2017-06-13 Amazon Technologies, Inc. Low latency computational capacity provisioning
DE102014221247A1 (de) * 2014-10-20 2016-04-21 Continental Automotive Gmbh Computersystem für ein Kraftfahrzeug
KR102301230B1 (ko) 2014-12-24 2021-09-10 삼성전자주식회사 가상화된 gpu들에 대한 스케줄링을 수행하는 방법 및 이를 위한 디바이스
CN104660711A (zh) * 2015-03-13 2015-05-27 华存数据信息技术有限公司 一种基于图形处理器虚拟化的远程可视化应用方法
CN105389199B (zh) * 2015-10-21 2019-09-27 同济大学 一种基于Xen的FPGA加速器虚拟化平台及应用
US9971708B2 (en) * 2015-12-02 2018-05-15 Advanced Micro Devices, Inc. System and method for application migration between docking station and dockable device
EP3188014B1 (en) * 2015-12-29 2022-07-13 Dassault Systèmes Management of a plurality of graphic cards
EP3188013B1 (en) * 2015-12-29 2022-07-13 Dassault Systèmes Management of a plurality of graphic cards
US10685473B2 (en) * 2017-05-31 2020-06-16 Vmware, Inc. Emulation of geometry shaders and stream output using compute shaders
US11029986B2 (en) 2018-05-25 2021-06-08 Microsoft Technology Licensing, Llc Processor feature ID response for virtualization
US10853115B2 (en) * 2018-06-25 2020-12-01 Amazon Technologies, Inc. Execution of auxiliary functions in an on-demand network code execution system
US11943093B1 (en) 2018-11-20 2024-03-26 Amazon Technologies, Inc. Network connection recovery after virtual machine transition in an on-demand network code execution system
US12327133B1 (en) 2019-03-22 2025-06-10 Amazon Technologies, Inc. Application gateways in an on-demand network code execution system
US11048544B2 (en) * 2019-04-08 2021-06-29 Sap Se Cloud resource credential provisioning for services running in virtual machines and containers
CN110442389B (zh) * 2019-08-07 2024-01-09 北京技德系统技术有限公司 一种多桌面环境共享使用gpu的方法
CN111913794B (zh) * 2020-08-04 2024-08-09 北京百度网讯科技有限公司 用于共用gpu的方法、装置、电子设备及可读存储介质
CN112991505B (zh) * 2021-04-16 2022-02-18 辽宁向日葵教育科技有限公司 一种线上3d渲染方法
US11968280B1 (en) 2021-11-24 2024-04-23 Amazon Technologies, Inc. Controlling ingestion of streaming data to serverless function executions
US12015603B2 (en) 2021-12-10 2024-06-18 Amazon Technologies, Inc. Multi-tenant mode for serverless code execution
CN115080183B (zh) * 2022-06-10 2024-12-06 阿里巴巴（中国）有限公司 一种vgpu加速方法、设备及存储介质
US12381878B1 (en) 2023-06-27 2025-08-05 Amazon Technologies, Inc. Architecture for selective use of private paths between cloud services
US12476978B2 (en) 2023-09-29 2025-11-18 Amazon Technologies, Inc. Management of computing services for applications composed of service virtual computing components

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060146057A1 (en) * 2004-12-30 2006-07-06 Microsoft Corporation Systems and methods for virtualizing graphics subsystems
US20070091102A1 (en) * 2005-10-26 2007-04-26 John Brothers GPU Pipeline Multiple Level Synchronization Controller Processor and Method
US20090322784A1 (en) * 2008-02-27 2009-12-31 Gabriele Sartori System and method for virtual 3d graphics acceleration and streaming multiple different video streams
US20110084973A1 (en) * 2009-10-08 2011-04-14 Tariq Masood Saving, Transferring and Recreating GPU Context Information Across Heterogeneous GPUs During Hot Migration of a Virtual Machine

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7154500B2 (en) 2004-04-20 2006-12-26 The Chinese University Of Hong Kong Block-based fragment filtration with feasible multi-GPU acceleration for real-time volume rendering on conventional personal computer
US7920144B2 (en) 2005-01-18 2011-04-05 Siemens Medical Solutions Usa, Inc. Method and system for visualization of dynamic three-dimensional virtual objects
US7394464B2 (en) 2005-01-28 2008-07-01 Microsoft Corporation Preshaders: optimization of GPU programs
US7928990B2 (en) 2006-09-27 2011-04-19 Qualcomm Incorporated Graphics processing unit with unified vertex cache and shader register file
US8065687B2 (en) 2007-01-05 2011-11-22 Moka5, Inc. Bypass virtualization
US7821521B2 (en) * 2007-02-28 2010-10-26 Red Hat, Inc. Methods and systems for legacy graphics emulation
US8310491B2 (en) 2007-06-07 2012-11-13 Apple Inc. Asynchronous notifications for concurrent graphics operations
JP5525175B2 (ja) * 2008-04-08 2014-06-18 アビッド テクノロジー インコーポレイテッド 複数のハードウェア・ドメイン、データ・タイプ、およびフォーマットの処理を統合し抽象化するフレームワーク
US8432405B2 (en) 2008-06-26 2013-04-30 Microsoft Corporation Dynamically transitioning between hardware-accelerated and software rendering
US8661355B1 (en) * 2008-10-10 2014-02-25 Cisco Technology, Inc. Distinguishing shared and non-shared applications during collaborative computing sessions
US8711159B2 (en) 2009-02-23 2014-04-29 Microsoft Corporation VGPU: a real time GPU emulator
US8237726B2 (en) 2009-06-26 2012-08-07 Intel Corporation Register allocation for message sends in graphics processing pipelines
US9245371B2 (en) 2009-09-11 2016-01-26 Nvidia Corporation Global stores and atomic operations
US8917271B2 (en) * 2009-10-05 2014-12-23 Nvidia Corporation Redistribution of generated geometric primitives
US20110102443A1 (en) * 2009-11-04 2011-05-05 Microsoft Corporation Virtualized GPU in a Virtual Machine Environment
JP5316432B2 (ja) * 2010-01-19 2013-10-16 富士通株式会社 ネットワークコントローラの制御方法、プログラム及び情報処理装置
US8970608B2 (en) 2010-04-05 2015-03-03 Nvidia Corporation State objects for specifying dynamic state
US8842122B2 (en) * 2011-12-15 2014-09-23 Qualcomm Incorporated Graphics processing unit with command processor

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060146057A1 (en) * 2004-12-30 2006-07-06 Microsoft Corporation Systems and methods for virtualizing graphics subsystems
US20070091102A1 (en) * 2005-10-26 2007-04-26 John Brothers GPU Pipeline Multiple Level Synchronization Controller Processor and Method
US20090322784A1 (en) * 2008-02-27 2009-12-31 Gabriele Sartori System and method for virtual 3d graphics acceleration and streaming multiple different video streams
US20110084973A1 (en) * 2009-10-08 2011-04-14 Tariq Masood Saving, Transferring and Recreating GPU Context Information Across Heterogeneous GPUs During Hot Migration of a Virtual Machine

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2018512661A (ja) * 2015-03-23 2018-05-17 インテル コーポレイション グラフィックス・プロセッサ仮想化のためのシャドー・コマンド・リング

Also Published As

Publication number Publication date
US20130181998A1 (en) 2013-07-18
CN104067230B (zh) 2017-09-08
EP2805231A4 (en) 2015-09-02
JP6140190B2 (ja) 2017-05-31
KR102013002B1 (ko) 2019-08-21
EP2805231A1 (en) 2014-11-26
CN104067230A (zh) 2014-09-24
JP2015511346A (ja) 2015-04-16
KR20140117414A (ko) 2014-10-07
US8941670B2 (en) 2015-01-27

Similar Documents

Publication Publication Date Title
US8941670B2 (en) Para-virtualized high-performance computing and GDI acceleration
US8941671B2 (en) Para-virtualized domain, hull, and geometry shaders
US8274518B2 (en) Systems and methods for virtualizing graphics subsystems
US8692832B2 (en) Para-virtualized asymmetric GPU processors
CN104737129B (zh) 用于对来宾虚拟机的硬件资源进行虚拟化的专用虚拟机
US11157302B2 (en) Idle processor management in virtualized systems via paravirtualization
US9442754B1 (en) Deferred asynchronous actions for virtual devices
JP2015511346A5 (enExample)
CN116113974A (zh) 生成插值视频帧的技术
KR20230002058A (ko) 동기화 장벽
CN112817739A (zh) 用于修改可执行图以实施不同工作负载的技术
EP3113015B1 (en) Method and apparatus for data communication in virtualized environment
WO2007024444A1 (en) Method and apparatus for supporting universal serial bus devices in a virtualized environment
US20230122396A1 (en) Enabling shared graphics and compute hardware acceleration in a virtual environment
WO2022221538A1 (en) Application programming interface to monitor resource usage

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13738213

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2013738213

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 20147019876

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2014553324

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE