Graphics processing on a processor core

• US 7,656,409 B2
• Filed: 12/23/2005
• Issued: 02/02/2010
• Est. Priority Date: 12/23/2005
• Status: Active Grant

• Current Assignee: Tahoe Research Limited (f/k/a Learndale Limited) (Vector Capital Corporation)
• Original Assignee: Intel Corporation
• Inventors: Rasheed, Yasser, Cool, Lyle
• Primary Examiner(s): GOOD JOHNSON, MOTILEWA
• Application Number:
  US11/317,829
• Publication Number:
  US 20070146373A1
• Time in Patent Office: 1,502 Days
• Field of Search: 345/502, 345/501, 345/522, 712/220, 712/225
• US Class Current: 345/501
• CPC Class Codes: G06F 9/5044 considering hardware capabilities

Abstract

In a many core system, receiving a call to a graphics driver; translating the call into a command executable on a core of the many core system; and executing the translated call on the core.

Claims

1. In a many core system, a method comprising:

receiving a call to a graphics driver in a many core system, the graphics driver managing execution of the call on an execution platform that includes a dedicated graphics processing unit (“

GPU”

) interconnected to the many core system and one or more cores in the many core system, the graphics driver;

executing the call on the dedicated GPU after determining that the dedicated GPU is available and able to execute the call and that an estimated performance of the call on the available and able dedicated GPU alone is sufficient;

executing the call on a single core of the many core system after determining that the dedicated GPU is unavailable or unable or that the estimated performance of the call on the available and able dedicated GPU is insufficient, and that the single core is available to execute the call and that an estimated performance of the call on the available single core alone is sufficient;

executing the call in parallel on one or both of the available and able dedicated GPU and one or more available cores of the many core system after determining that the estimated performance of the execution of the call on either the available single core alone or the available and able dedicated GPU alone is insufficient; and

translating the call into a command executable on an available core of the many core system when executing the call on the available core.

2. The method of claim 1 wherein translating the call into a command executable on the core of the many core system farther comprises a translation layer of the system performing the translation of the call into a command executable on the core of the many core system.

3. The method of claim 2 further comprising

the translation layer partitioning the execution of the command into a plurality of threads;

and

the translation layer scheduling the execution of the plurality of threads on at least the graphics processing unit, and the core.

4. The method of claim 2 wherein the many core system further comprises a plurality of cores, further comprising the translation layer partitioning the execution of the command into a plurality of threads and the translation layer scheduling the execution of the plurality of threads on at least the graphics processing unit and one or more of the plurality of cores.

5. The method of claim 1 wherein the core further comprises a segregated core of the many core system configured to be inaccessible to the operating system.

6. The method of claim 5 wherein a firmware layer provides an interface to the segregated core;

and wherein executing the translated call on the core further comprises using a service of the firmware layer to execute the translated call on the segregated core.

7. A machine-readable medium having stored thereon data that when accessed via machine causes the machine to perform a method comprising:

receiving a call to a graphics driver in a many core system, the graphics driver managing execution of the call on an execution platform that includes a dedicated graphics processing unit (“

GPU”

) interconnected to the many core system and one or more cores in the many core system, the graphics driver;

executing the call on the dedicated GPU after determining that the dedicated GPU is available and able to execute the call and that an estimated performance of the call on the available and able dedicated GPU alone is sufficient;

executing the call on a single core of the many core system after determining that the dedicated GPU is unavailable or unable or that the estimated performance of the call on the available and able dedicated GPU is insufficient, and that the single core is available to execute the call and that an estimated performance of the call on the available single core alone is sufficient;

executing the call in parallel on one or both of the available and able dedicated GPU and one or more available cores of the many core system after determining that the estimated performance of the execution of the call on either the available single core alone or the available and able dedicated GPU alone is insufficient; and

translating the call into a command executable on an available core of the many core system when executing the call on the available core.

8. The machine-readable medium of claim 7 wherein translating the call into a command executable on the core of the many core system further comprises a translation layer of the system performing the translation of the call into a command executable on the core of the many core system.

9. The machine-readable medium of claim 8 farther comprising

the translation layer partitioning the execution of the command into a plurality of threads;

and

the translation layer scheduling the execution of the plurality of threads on at least the graphics processing unit, and the core.

10. The machine-readable medium of claim 8 where in the many core system further comprises a plurality of cores, further comprising the translation layer partitioning the execution of the command into a plurality of threads and the translation layer scheduling the execution of the plurality of threads on at least the graphics processing unit and one or more of the plurality of cores.

11. The machine-readable medium of claim 7 wherein the core farther comprises a segregated core of the many core system configured to be inaccessible to the operating system.

12. The machine-readable medium of claim 11 wherein a firmware layer provides an interface to the segregated core;

and wherein executing the translated call on the core further comprises using a service of the firmware layer to execute the translated call on the segregated core.

13. A many core system comprising:

a plurality of cores in the many core system;

a memory including a graphics driver module; and

a store;

wherein the graphics driver module is adapted to receive a call to a graphics driver, the graphics driver further adapted to manage execution of the call on an execution platform that includes a dedicated graphics process unit (“

GPU”

) interconnected to the many core system, including being adapted to;

execute the call on the dedicated GPU after determining that the dedicated GPU is available and able to execute determine whether the call and that an estimated performance of the call on is executable either on the core or on the available and able dedicated graphics processing unit GPU alone is sufficient;

execute the call on a single core of the many core system after determining that the dedicated GPU is unavailable or unable or that the estimated performance of the call on the available and able dedicated GPU is insufficient, and that the single core is available to execute the call and that an estimated performance of the call on the available single core alone is sufficient;

execute the call in parallel on one or both of the available and able dedicated GPU and one or more available cores of the many core system after determining that the estimated performance of the execution of the call on either the available single core alone or the available and able dedicated GPU alone is insufficient; and

wherein the graphics driver module is further adapted to translate the call into a command executable on an available core of the plurality of cores; and

to execute the translated call on the available core of the plurality of cores.

14. The system of claim 13 wherein the one of the plurality of cores further comprises a segregated core of the many core system configured to be inaccessible to the operating system.

15. The system of claim 14 further comprising a firmware layer to provide an interface to the segregated core;

and to provide a service to execute the translated call on the segregated core.

16. The system of claim 13 wherein the memory further comprises dynamic random access memory (DRAM).