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. A processor comprising:
a first central processing unit (CPU);
a second central processing unit;
a non-CPU functional unit;
at least one cache associated with the non-CPU functional unit that can not be snooped by the first central processing unit or by the second central processing unit; and
interconnect logic coupled to the first central processing unit, the second central processing unit, and the non-CPU functional unit;
wherein the interconnect logic enables information sharing between the first central processing unit, the second central processing unit, and the non-CPU functional unit without first resorting to main memory.
2. The processor of claim 1 , wherein the non-CPU functional unit is a graphics processing unit.
3. The processor of claim 2 , wherein the interconnect logic comprises snoop logic to enable the graphics processing unit to access a cache associated with the first central processing unit.
4. The processor of claim 3 , wherein the cache associated with the first central processing unit is a last-level-cache.
5. The processor of claim 4 , wherein the last-level-cache is an L2 cache.
6. The processor of claim 2 , further comprising translation logic coupled to the graphics processing unit to translate a virtual address to a physical address in response to a memory access by the graphics processing unit.
7. The processor of claim 2 , wherein said at least one cache associated with the non-CPU functional unit that can not be snooped by the first central processing unit or by the second central processing unit is a portion of a cache hierarchy associated with the graphics processing unit that is virtually addressed by the graphics processing unit.
8. The processor of claim 2 , the first central processing unit having at least an associated L1 cache, wherein the interconnect logic comprises snoop logic to enable the graphics processing unit to have access to information modifiable by the first central processing unit by snooping the L1 cache associated with the first central processing unit.