Technique to share information among different cache coherency domains
A technique to enable information sharing among agents within different cache coherency domains. In one embodiment, a graphics device may use one or more caches used by one or more processing cores to store or read information, which may be accessed by one or more processing cores in a manner that does not affect programming and coherency rules pertaining to the graphics device.
1. An apparatus comprising:
a first cache and a second cache within a first coherency domain, wherein the first cache is at a lower level in a first cache hierarchy than the second cache; and
a processor corresponding to a second cache hierarchy including a third cache, wherein the first cache hierarchy is in a different coherency domain than a portion of the second cache hierarchy such that the processor can not snoop at least one cache in said first cache hierarchy and wherein the processor is to snoop at least one of the first and second caches to enable information sharing between the processor and the first coherency domain without first resorting to main memory.
2. The apparatus of claim 1 , said second cache hierarchy including a fourth cache, wherein the third cache is at a lower level in the second cache hierarchy than the fourth cache.
3. The apparatus of claim 1 , wherein said processor is coupled with a system memory via a memory controller.
4. The apparatus of claim 1 , one of said first coherency domain and said different coherency domain comprising a graphics coherency domain.
5. The apparatus of claim 1 , comprising a graphics processor.
6. The apparatus of claim 1 , comprising a central processing unit.
7. The apparatus of claim 1 , comprising an I/O device.
8. A system comprising:
a first cache and a second cache within a first coherency domain, wherein the first cache is at a lower level in a first cache hierarchy than the second cache;
a processor corresponding to a second cache hierarchy including a third cache; and
a system memory coupled with said processor, wherein the first cache hierarchy is in a different coherency domain than a portion of the second cache hierarchy such that the processor can not snoop at least one cache in the first cache hierarchy and wherein the processor is to snoop at least one of the first and second caches to enable information sharing between the processor and the first coherency domain without requiring an access to the system memory.
9. The system of claim 8 , said first coherency domain and different coherency domain comprising a graphics coherency domain.
10. The system of claim 9 , wherein said processor is coupled with the system memory via a memory controller.
11. The system of claim 10 , the processor comprising a central processing unit.
12. The system of claim 10 , said second cache hierarchy including a fourth cache, wherein the third cache is at a lower level in the second cache hierarchy than the fourth cache.
13. The system of claim 8 , comprising an I/O device.
14. A computer-implemented method comprising:
storing data from a first cache to a second cache within a first coherency domain, wherein the first cache is at a lower level in a first cache hierarchy than the second cache; and
issuing a snoop from a processor corresponding to a second cache hierarchy to at least one of the first and second caches, the second cache hierarchy including a third cache, wherein the first cache hierarchy is in a different coherency domain than a portion of the second cache hierarchy such that the processor can not snoop at least one cache in the first cache hierarchy.
15. The method of claim 14 , said second cache hierarchy including a fourth cache, wherein the third cache is at a lower level in the second cache hierarchy than the fourth cache.
16. The method of claim 14 , said first coherency domain and different coherency domain comprising a graphics coherency domain.
17. The method of claim 14 , wherein said processor is coupled with the system memory via a memory controller.
18. The method of claim 14 , wherein said processor is a central processing unit.
19. The method of claim 14 , one of said first cache hierarchy and said second cache hierarchy corresponding to a graphics processor.
20. The method of claim 19 , the other one of said first cache hierarchy and said second cache hierarchy corresponding to a central processing unit.