# US 2010/0146216 A1 · Method and apparatus for filtering memory write snoop activity in a distributed shared memory computer

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0146216 A1
- Document ID: US20100146216A1
- Status: Published application
- Kind: A1
- Application number: 12/708,013
- Filed: 2010-02-18
- Published: 2010-06-10
- Earliest claimed: 2004-04-07
- Assignees: Globalfoundries Inc.
- Inventors: Patrick N. Conway
- Web page: https://patentgrep.com/patent/US20100146216A1

## CPC

- G06F 12/082: {Associative directories (G06F12/0822 takes precedence)}
- Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

A method and apparatus for filtering memory probe activity for writes in a distributed shared memory computer. In one embodiment, the method may include assigning an uncached directory state to a cache data block in response to evicting the cache data block. In another embodiment, the method may include assigning a remote directory state to a cache data block in response to evicting the cache data block and storing it in a remote cache. In a third embodiment, the method may include assigning a pairwise-shared directory state in response to a second processor node initiating a load operation to a cache data block in a modified cache state in a first processor node. In a fourth embodiment, the method may include assigning a migratory directory state in response to a processor node initiating a store operation to a cache data block in a pairwise-shared cache state.

## Claims

### Claim at position 1

1-43. (canceled)

### Claim 44

A method comprising:
initiating a first store operation to a cache data block from a first processing node;
initiating a first load operation to said cache data block from a second processing node subsequent to initiating said first store operation;
assigning a pairwise-shared directory state to a coherence directory entry corresponding to said cache data block in response to initiating said first load operation;
initiating a second store operation to said cache data block from said second processing node subsequent to initiating said first load operation; and
assigning a migratory directory state to said coherence directory entry in response to initiating said second store operation.

### Claim 45 (depends on 44)

The method as recited in claim 44, wherein said migratory directory state is distinct from a modified directory state.

### Claim 46 (depends on 44)

The method as recited in claim 44, wherein assigning a migratory directory state further comprises storing an identity of said second processing node.

### Claim 47 (depends on 44)

The method as recited in claim 44, further comprising issuing a probe only to said first processing node in response to initiating said second store operation.

### Claim 48 (depends on 44)

The method as recited in claim 44, further comprising:
initiating a second load operation to said cache data block from a third processing node subsequent to assigning a migratory directory state to said coherence directory entry;
retaining said migratory directory state assigned to said coherence directory entry in response to initiating said second load operation;
issuing an invalidating probe only to said second processing node in response to initiating said second load operation; and
assigning a modified cache state to said cache data block in response to initiating said second load operation.

### Claim 49

A system comprising:
a coherence directory;
a first processing node coupled to said coherence directory and comprising a first cache; and
a second processing node coupled to said coherence directory;
wherein said first processing node is configured to initiate a first store operation to a cache data block;
wherein said second processing node is configured to initiate a first load operation to said cache data block subsequent to said first store operation;
wherein said coherence directory is configured to transition a coherence directory entry corresponding to said cache data block to a pairwise-shared directory state in response to said second processing node initiating said first load operation to said cache data block;
wherein said second processing node is configured to initiate a second store operation to said cache data block subsequent to said first load operation;
wherein said coherence directory is configured to transition said coherence directory entry of said cache data block from a pairwise-shared directory state to a migratory directory state in response to said second processing node initiating said second load operation.

### Claim 50 (depends on 49)

The system as recited in claim 49, wherein said migratory directory state is distinct from a modified directory state.

### Claim 51 (depends on 49)

The system as recited in claim 49, wherein transitioning to a migratory directory state further comprises and storing an identity of said second processing node.

### Claim 52 (depends on 49)

The system as recited in claim 49, wherein said coherence directory is further configured to cause a probe to be issued only to said first processing node in response to said second processing node initiating said second store operation.

### Claim 53 (depends on 49)

The system as recited in claim 49, further comprising a third processing node coupled to said first and said second processing nodes, wherein said third processing node is configured to initiate a second load operation to said cache data block subsequent to said second store operation, wherein said coherence directory is configured to retain said migratory directory state of said coherence directory entry in response to said third processing node initiating said second load operation, and wherein said coherence directory is further configured to cause a probe to be issued only to said second processing node in response to said third processing node initiating said second load operation.
