Apparatus and method for maintaining data coherence within a cluster of symmetric multiprocessors

• US 6,341,339 B1
• Filed: 03/26/1998
• Issued: 01/22/2002
• Est. Priority Date: 03/26/1998
• Status: Expired due to Fees

• Current Assignee: Compaq Information Technologies Group LP (HP Inc.), University of Rochester
• Original Assignee: Compaq Computer Corporation (HP Inc.)
• Inventors: Scott, Michael L., Hunt, Galen, Stets, Robert, Hardavellas, Nikos, Dwarkadas, Sandhya, Kontothanassis, Leonidas
• Primary Examiner(s): Nguyen, Hiep T.
• Assistant Examiner(s): MCLEAN MAYO, KIMBERLY N
• Application Number:
  US09/048,232
• Time in Patent Office: 1,398 Days
• Field of Search: 711/141, 711/117, 711/112, 711/113, 711/114, 711/147, 711/148, 709/203, 709/201, 709/202, 709/216, 709/217, 709/400, 707/201, 707/203, 707/204, 707/202
• US Class Current: 711/148
• CPC Class Codes: G06F 12/0813 with a network or matrix configuration
• CPC Class Codes: G06F 12/0837 with software control, e.g. non-cacheable data
• CPC Class Codes: G06F 9/544 Buffers; Shared memory; Pipes
• CPC Class Codes: Y10S 707/99952 Coherency, e.g. same view to multiple users

Abstract

The present invention discloses an apparatus and method for maintaining the coherence of data within a shared memory network including a plurality of nodes. The system utilizes processors monitoring the occurrence of particular processing events within a local memory storage area. Upon the detection of events indicating the change of status of a particular group of data, a comparison is made between a modified copy of the group of data and a clean copy of the group of data to detect any modifications made to the group of data. These modifications are entered into the clean copy of the group of data and processing continues.

Claims

1. A method for maintaining coherence of data within a shared memory network, comprising the steps of:

providing a home node copy of a group of data at a first electronic node, the home node copy representing a master copy of the group of data generated at the first electronic node;

storing a working copy of at least a first portion of the master copy at a second electronic node;

storing a twin copy of the working copy at the second electronic node;

fetching the master copy from the first electronic node;

determining a difference between the fetched master and the twin copy;

writing the determined difference to the working copy; and

writing the determined difference to the twin copy.

2. The method of claim 1, further comprising the step of:

detecting initiation of a fetch operation of the home node copy of the group of data.

3. The method of claim 2, further comprising the step of:

modifying the working copy of data;

wherein the step of storing the twin copy includes the step of storing the twin copy responsive to the step of modifying the working copy.

4. The method of claim 1, wherein the difference determination step further comprises the step of comparing the home node copy to the twin copy using an exclusive OR operation at the bit level.

8. The method of claim 1, wherein the modification of the at least a first portion of the group of data is preceded by a write instruction.

5. An article of manufacture, comprising:

a processor readable storage medium having processor programming stored thereon for maintaining coherence of data within a shared memory network including a plurality of processors, wherein said processor programming is configured to be readable from at least one of said plurality of processors causing said at least one processor to;

fetch a master copy of data from a first electronic node;

store a working copy of the master copy of data at a second electronic node;

store a twin copy of the working copy the second electronic node;

determine a difference between the fetched master copy and the stored twin copy;

write the determined difference to the working copy; and

write the determined difference to the twin copy.

6. The article of manufacture of claim 5, wherein the processor programming is further configured to cause said processor to:

detect initiation of a fetch operation of the master copy of data.

7. An article of manufacture, comprising:

a processor readable storage medium having processor programming stored thereon for maintaining coherence of data within a shared memory network including a plurality of processors, wherein said processor programming is configured to be readable from at least one of said plurality of processors causing said at least one processor to;

fetch a master copy of data from a first electronic node;

detect initiation of a fetch operation of the master copy of data;

store a working copy of the master copy at a second electronic node;

store a twin copy of the working copy at the second electronic node;

modify at least a portion of the working copy;

store the twin copy responsive to the step of modifying the working copy;

determine a difference between the fetched master copy and the stored twin copy;

write the determined difference to the working copy; and

write the determined difference to the twin copy.

9. A method for maintaining coherence of data within a shared memory network, comprising the steps of:

providing a master copy of a group of data at a first node of the network;

storing a working copy of at least a first portion of the master copy at a second node of the network;

storing a twin copy of the working copy at the second node of the network, the twin copy being created when the working copy has been modified;

fetching the master copy of the group of data from the first node of the network;

determining a difference between the fetched master copy of the group of data and the stored twin copy of the group of data;

writing the determined difference to the working copy and

writing the determined difference to the twin copy.

10. The method of claim 9, wherein the twin copy is created upon detection of a write instruction with respect to the working copy of data.

11. The method of claim 9, further comprising the step of detecting initiation of a fetch operation of the master copy of data.

12. The method of claim 11, further comprising the step of modifying the working copy of data;

wherein the step of storing the twin copy includes the step of storing the twin copy responsive to the step of modifying the working copy.

13. The method of claim 9, wherein the difference determination step further comprises the step of performing a bit by bit comparison of the master copy of the group of data to a twin copy of data using an exclusive-OR operation.