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 system in a processor comprising:
a central processing unit (CPU) within a CPU cache coherency domain having a first set of coherency rules;
a graphics device within a graphics cache coherency domain having a second set of coherency rules different from the first set of coherency rules;
a CPU cache hierarchy including the CPU and including:
a level 1 (L1) CPU cache operatively coupled with the CPU, and
a second cache at a different level than the L1 CPU cache, the second cache being physically addressed; and
a graphics cache hierarchy including the graphics device and including:
an L1 graphics cache operatively coupled with the graphics device, and
a third cache at a different level than the L1 graphics cache, the third cache being physically addressed;
said CPU to apply the first set of coherency rules to the third cache, and said graphics device to apply the second set of coherency rules to the third cache, thereby allowing data to be shared between the CPU cache coherency domain and the graphics cache coherency domain.
2. The system of claim 1 wherein the third cache is a mid-level cache (MLC).
3. The system of claim 2 wherein the second cache is a last-level cache (LLC) inclusive of information stored in each of the third cache and the L1 CPU cache.
4. The system of claim 1 wherein said graphics device is to store data from the L1 graphics cache to the third cache in response to a graphics rendering event.
5. The system of claim 4 wherein the L1 graphics cache is virtually addressed.
6. The system of claim 4 wherein said CPU is to snoop the third cache thereby allowing data to be shared between the CPU cache coherency domain and the graphics cache coherency domain.
7. The system of claim 4 wherein said CPU is to apply the first set of coherency rules to all physically addressed structures, including the second and third caches and main memory, thereby allowing data to be shared between the CPU cache coherency domain and the graphics cache coherency domain.