Private memory table for reduced memory coherence traffic

1. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device comprising a multiple processor system having multiple processor chips with each processor chip having one or more processing cores, causes the computing device to:

create, for each processing core of each processor chip, an associated processor core level private region table;

create, for each processor chip, an associated chip level private region table, wherein entries in each processor core level private region table and chip level private region table correspond to remote memory regions that are private to a virtual machine executing on a processor core associated with the processor core level region table and associated with a chip corresponding to the chip level private region table;

responsive to a memory operation from a processing core in a first processing node of the multiple processor system resulting in a cache miss, check a processor core level private region table associated with the processing core, wherein the memory operation attempts to access a memory region stored in memory of a second processing node of the multiple processor system; and

responsive to determining the memory region corresponds to an entry in the processor core level private region table and the memory region is node-contained in the second processing node, perform the memory operation on the second processing node without snooping the entire multiple processor system.

2. The computer program product of claim 1 , wherein the computer readable program further causes the computing device to perform the memory operation using snoopy protocol responsive to determining the memory region does not correspond to an entry in the private region table.

3. The computer program product of claim 1 , wherein the computer readable program further causes the computing device to perform the memory operation using snoopy protocol responsive to determining the memory region is not node-contained in the second processing node.

4. The computer program product of claim 1 , wherein each entry of the processor core level private region table comprises a tag, a size, least recently used (LRU) information, and a count of cache lines stored in a cache of the processing core.

5. The computer program product of claim 1 , wherein the computer readable program further causes the computing device to:

responsive to migrating a virtual machine executing on the processing core to a new processing core, migrate entries of the private region table to a private region table associated with the new processing core.

6. The computer program product of claim 1 , wherein each entry in the chip level private region table comprises a tag, a size, least recently used (LRU) information, a count of cache lines stored in a cache of the processing core, and a processor core identifier that identifies a processor core associated with a memory region corresponding to the entry.

7. The computer program product of claim 6 , wherein the chip level private region table stores entries that are cast out of processor core level private region tables associated with processor cores of the chip, and wherein the processor core identifier identifies which processor core of the chip from which the corresponding entry in a processor core level private region table was cast out and stored in the chip level private region table.

8. An apparatus comprising:

a processor, wherein the processor comprises multiple processor chips with each processor chip having one or more processing cores; and

a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to:

create, for each processing core of each processor chip, an associated processor core level private region table;

create, for each processor chip, an associated chip level private region table, wherein entries in each processor core level private region table and chip level private region table correspond to remote memory regions that are private to a virtual machine executing on a processor core associated with the processor core level region table and associated with a chip corresponding to the chip level private region table;

responsive to a memory operation from a processing core in a first processing node of the multiple processor system resulting in a cache miss, check a processor core level private region table associated with the processing core, wherein the memory operation attempts to access a memory region stored in memory of a second processing node of the multiple processor system; and

responsive to determining the memory region corresponds to an entry in the processor core level private region table and the memory region is node-contained in the second processing node, perform the memory operation on the second processing node without snooping the entire multiple processor system.

9. The apparatus of claim 8 , wherein the instructions further cause the processor to perform the memory operation using snoopy protocol responsive to determining the memory region does not correspond to an entry in the private region table.

10. The apparatus of claim 8 , wherein the instructions further cause the processor to perform the memory operation using snoopy protocol responsive to determining the memory region is not node-contained in the second processing node.

11. The apparatus of claim 8 , wherein each entry of the processor core level private region table comprises a tag, a size, least recently used (LRU) information, and a count of cache lines stored in a cache of the processing core.

12. The apparatus of claim 8 , wherein each entry in the chip level private region table comprises a tag, a size, least recently used (LRU) information, a count of cache lines stored in a cache of the processing core, and a processor core identifier that identifies a processor core associated with a memory region corresponding to the entry.

13. The apparatus of claim 12 , wherein the chip level private region table stores entries that are cast out of processor core level private region tables associated with processor cores of the chip, and wherein the processor core identifier identifies which processor core of the chip from which the corresponding entry in a processor core level private region table was cast out and stored in the chip level private region table.

14. A method, in a multiple processor system having multiple processor chips with each processor chip having one or more processing cores, the method comprising:

creating, for each processing core of each processor chip, an associated processor core level private region table;

creating, for each processor chip, an associated chip level private region table, wherein entries in each processor core level private region table and chip level private region table correspond to remote memory regions that are private to a virtual machine executing on a processor core associated with the processor core level region table and associated with a chip corresponding to the chip level private region table;

checking, responsive to a memory operation from a processing core in a first processing node of the multiple processor system resulting in a cache miss, a processor core level private region table associated with the processing core, wherein the memory operation attempts to access a memory region stored in memory of a second processing node of the multiple processor system; and

performing, responsive to determining the memory region corresponds to an entry in the processor core level private region table and the memory region is node-contained in the second processing node, the memory operation on the second processing node without snooping the entire multiple processor system.

15. The method of claim 14 , further comprising performing the memory operation using snoopy protocol responsive to determining the memory region does not correspond to an entry in the private region table.

16. The method of claim 14 , further comprising performing the memory operation using snoopy protocol responsive to determining the memory region is not node-contained in the second processing node.

17. The method of claim 14 , wherein each entry of the processor core level private region table comprises a tag, a size, least recently used (LRU) information, and a count of cache lines stored in a cache of the processing core.

18. The method of claim 14 , wherein the computer readable program further causes the computing device to:

responsive to migrating a virtual machine executing on the processing core to a new processing core, migrate entries of the private region table to a private region table associated with the new processing core.

19. The method of claim 14 , wherein each entry in the chip level private region table comprises a tag, a size, least recently used (LRU) information, a count of cache lines stored in a cache of the processing core, and a processor core identifier that identifies a processor core associated with a memory region corresponding to the entry.

20. The method of claim 19 , wherein the chip level private region table stores entries that are cast out of processor core level private region tables associated with processor cores of the chip, and wherein the processor core identifier identifies which processor core of the chip from which the corresponding entry in a processor core level private region table was cast out and stored in the chip level private region table.