SHARED BUFFER TECHNIQUES FOR HETEROGENEOUS HYBRID GRAPHICS

• US 20110157189A1
• Filed: 12/31/2009
• Published: 06/30/2011
• Est. Priority Date: 12/31/2009
• Status: Active Grant

• Current Assignee: NVIDIA Corporation
• Original Assignee: NVIDIA Corporation
• Inventors: Diard, Franck
• Granted Patent: US 9,111,325 B2
• Time in Patent Office: Days
• US Class Current: 345/502
• CPC Class Codes: G06T 1/00 General purpose image data processing

Abstract

The graphics processing technique includes detecting a transition from rendering graphics on a first graphics processing unit to a second graphics processing, by a hybrid driver. The hybrid driver, in response to detecting the transition, configures the first graphics processing unit to create a frame buffer. Thereafter, an image rendered on the second graphics processing unit may be copied to the frame buffer of the first graphics processing unit. The rendered image in the frame buffer may then be scanned out on the display.

Claims

1. One or more computing device readable media having computing device executable instructions which when executed perform a method comprising:

detecting, by a device kernel mode driver agent, a transition from rendering graphics on a first graphics processing unit to a second graphics processing unit;

configuring, through the device kernel mode driver agent, the first graphics processing unit to create a frame buffer and to scan out to a display from the frame buffer, in response to detecting the transition; and

blitting an image rendered on the second graphics processing unit to the frame buffer of the first graphics processing unit after detecting the transition.

2. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 1, wherein the transition is initiated by a service or user interface calling a given entry point of a local function in the device kernel mode driver agent.

3. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 1, wherein configuring the first graphics processing unit to create the frame buffer comprise:

redirecting a call to allocate a surface from a kernel mode driver of the second graphics processing unit to a kernel mode driver of the first graphics processing unit;

allocating, by the first graphics processing unit, the frame buffer for a surface of a specified resolution in response to the redirected call to allocate the surface;

returning, from the first graphics processing unit to the device kernel mode driver agent, a virtual address of the frame buffer; and

determining, by the device kernel mode driver agent, a physical address of the frame buffer from the virtual address returned from the first graphics processing unit; and

storing, by the device kernel mode driver agent, the physical address of the frame buffer

4. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 3, wherein the device kernel mode driver agent includes the device specific kernel mode driver of the first graphics processing unit and the device specific kernel mode driver of the second graphics processing unit.

5. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 3, wherein the device kernel mode driver agent includes components of the device specific kernel mode driver of the first graphics processing unit and the device specific kernel mode driver of the second graphics processing unit.

6. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 3, wherein blitting the rendered image comprises:

receiving, by the device kernel mode driver agent, a call to the device specific kernel mode driver of the second graphics processing unit to display the rendered surface;

inserting, by the device kernel mode driver agent, a call to the device specific kernel mode driver of the second graphics processing unit to copy the rendered image to the physical address of the frame buffer of the first graphics processing unit; and

inserting, by the device kernel mode driver agent, a call to the device specific kernel mode driver of the first graphics processing unit to present the rendered image from the frame buffer, by a digital-to-analog converter of first graphics processing unit, on the display.

7. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 6, wherein copying the rendered image to the physical address of the frame buffer of the first graphics processing unit comprises a direct memory access (DMA).

8. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 1, wherein the first graphics processing unit comprises an integrated graphics processing unit.

9. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 1, wherein the second graphics processing unit comprises a discrete graphics processing unit.

10. The one or more computing device readable media having computing device executable instructions which when executed perform the method of claim 1, wherein the device kernel mode driver agent comprises a wrapper driver.

11. A method comprising:

loading a kernel mode wrapper driver that includes a kernel mode device driver for a first graphics processing unit attached to a primary display and a kernel mode device driver for a second graphics processing unit on an unattached adapter;

receiving, by the kernel mode wrapper driver, a call from a service or user interface to transition from rendering graphics on the first graphics processing unit to the second graphics processing unit;

outputting, by the kernel mode wrapper driver, a call to create a surface in a frame buffer for the first graphics processing unit in response to the received call to transition from rendering graphics on the first graphics processing unit to the second graphics processing unit;

receiving, by the kernel mode wrapper driver, a virtual address of the surface in the frame buffer for the first graphics processing unit in response to the call to create the surface;

determining, by the kernel mode wrapper driver, a physical address of the surface from the virtual address of the surface; and

storing, by the kernel mode wrapper driver, the physical address of the surface.

12. The method according to claim 11, further comprising:

determining, by the kernel mode wrapper driver, a context of one or more calls after the transition call;

receiving, by the kernel mode wrapper driver, one or more calls having a rendering context after the transition call; and

directing, by the kernel mode wrapper driver, the one or more calls having the rendering context to the kernel mode driver of the second graphics processing unit after the transition call.

13. The method according to claim 12, further comprising:

receiving, by the kernel mode wrapper driver, one or more calls having a display context after the transition call; and

redirecting, by the kernel mode wrapper driver, the one or more calls having the display context after the transition call.

14. The method according to claim 13, wherein redirecting the one or more calls having the display context further comprises:

receiving, by the kernel mode wrapper driver, a call to present an image rendered on the second graphics processing unit;

inserting, by the kernel mode wrapper driver, a call to copy the rendered image to the surface in the frame buffer for the first graphics processing unit at the physical address stored by the kernel mode wrapper driver; and

redirecting, by the kernel mode wrapper driver, a call to present the image from the surface in the frame buffer for the first graphics processing in response to the call to present the image rendered on the second graphics processing unit.

15. The method according to claim 14, wherein the call to copy the rendered image comprises a direct memory access to the physical address.

16. The method according to claim 14, wherein the rendered image is copied from a scanout of the second graphics processing unit to the surface in the frame buffer for the first graphics processing unit at the physical address.

17. The method according to claim 14, wherein the frame buffer is shared with the first and second graphics processing unit.

18. The method according to claim 14, wherein the rendered image is copied through a Peripheral Component Interconnect Express (PCIe) bus coupling the second graphics processing unit to the first graphics processing unit.

19. The method according to claim 14, wherein the frame buffer is shared with the first and second graphics processing unit and is mapped to a Peripheral Component Interconnect Express (PCIe) bus coupling the second graphics processing unit to the first graphics processing unit.

20. The method according to claim 14, wherein the frame buffer is double buffered.