Request and combined response broadcasting to processors coupled to other processors within node and coupled to respective processors in another node

• US 7,627,738 B2
• Filed: 12/19/2007
• Issued: 12/01/2009
• Est. Priority Date: 09/27/2005
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chung, Vicente E., Starke, William J., Goodman, Benjiman L., Reddy, Praveen S.
• Primary Examiner(s): Kim; Kenneth S
• Application Number:
  US11/960,283
• Publication Number:
  US 20080162872A1
• Time in Patent Office: 713 Days
• Field of Search: None
• US Class Current: 712/28
• CPC Class Codes: G06F 15/17337 Direct connection machines, e.g. completely connected computers, point to point communication networks coupling between buses G06F13/4004

Abstract

A data processing system includes a first processing node and a second processing node. The first processing node includes a plurality of first processing units coupled to each other for communication, and the second processing node includes a plurality of second processing units coupled to each other for communication. Each of the plurality of first processing units is coupled to a respective one of the plurality of second processing units in the second processing node by a respective one of a plurality of point-to-point links.

Claims

1. A data processing system, comprising:

a first processing node and a second processing node, wherein;

said first processing node includes a plurality of first processing units coupled to each other for communication and said second processing node includes a plurality of second processing units coupled to each other for communication;

said plurality of first processing units includes at least first, second and third processing units and said plurality of second processing units includes at least fourth, fifth and sixth processing units;

said first and fourth processing units are connected by a first point-to-point link;

said second and fifth processing units are connected by a second point-to-point link;

said third and sixth processing units are connected by a third point-to-point link;

said plurality of first processing units includes a node master processing unit and at least one node leaf processing unit;

said plurality of second processing units includes a remote hub processing unit and at least one remote leaf processing unit;

said node master processing unit broadcasts a request to each node leaf processing unit and to said remote hub processing unit;

said remote hub processing unit broadcasts said request to each remote leaf processing unit; and

said node master processing unit broadcasts a combined response for said request to each node leaf processing unit, remote hub processing unit and remote leaf processing unit based upon partial responses for said request received by said node master processing unit.

2. The data processing system of claim 1, wherein:

said first, second and third point-to-point links comprise a plurality of point-to-point second tier links; and

each of said plurality of first processing units is coupled to each other of said plurality of first processing units by a respective one of a plurality of point-to point first tier links.

3. The data processing system of claim 2, wherein each of said plurality of first processing units includes interconnect logic that broadcasts operations to all of said plurality of first processing units and all of said plurality of second processing units via all of said plurality of first tier links and one of said plurality of second tier links.

4. A data processing system, comprising:

a first processing node and a second processing node, wherein;

said first processing node includes a plurality of first processing units coupled to each other for communication and said second processing node includes a plurality of second processing units coupled to each other for communication;

said plurality of first processing units includes at least first, second and third processing units and said plurality of second processing units includes at least fourth, fifth and sixth processing units;

said first and fourth processing units are connected by a first point-to-point link;

said second and fifth processing units are connected by a second point-to-point link;

said third and sixth processing units are connected by a third point-to-point link;

wherein;

operations of said plurality of first and second processing units include, in order, at least a request phase in which a request is broadcast, a partial response phase in which individual processing units determine their respective responses to said request, and a combined response phase in which a system-wide combined response to said request is distributed; and

said plurality of first and second processing units route said combined response via each link traversed by said request in a same direction as said request and route at least one partial response via each link traversed by said request in an opposite direction to said request.

5. A first processing node for a data processing system including at least said first processing node and a second processing node, said first processing node comprising:

a plurality of first processing units coupled to each other for communication, wherein each of said plurality of first processing units includes interconnect logic by which each said first processing unit can be coupled to a respective one of a plurality of second processing units in said second processing node by a respective one of a plurality of point-to-point links, such that;

a first processing unit in said first processing node is coupled to a fourth processing unit in the second processing node by a first point-to-point link;

a second processing unit in said first processing node is coupled to a fifth processing unit in the second processing node by a second point-to-point link; and

a third processing unit in said first processing node is coupled to a sixth processing unit in the second processing node by a third point-to-point link;

wherein;

said plurality of first processing units includes a node master processing unit and at least one node leaf processing unit;

said plurality of second processing units includes a remote hub processing unit and at least one remote leaf processing unit;

said node master processing unit broadcasts a request to each node leaf processing unit and to said remote hub processing unit;

said node master processing unit broadcasts a combined response for said request to each node leaf processing unit, remote hub processing unit and remote leaf processing unit based upon partial responses for said request received by said node master processing unit.

6. The first processing node of claim 5, wherein:

said plurality of point-to-point links comprises a plurality of point-to-point second tier links; and

each of said plurality of first processing units is coupled to each other of said plurality of first processing units by a respective one of a plurality of point-to point first tier links.

7. The first processing node of claim 6, wherein each of said plurality of first processing units includes interconnect logic that broadcasts operations to all of said plurality of first processing units and all of said plurality of second processing units via all of said plurality of first tier links and one of said plurality of second tier links.

8. A first processing node for a data processing system including at least said first processing node and a second processing node, said first processing node comprising:

a plurality of first processing units coupled to each other for communication, wherein each of said plurality of first processing units includes interconnect logic by which each said first processing unit can be coupled to a respective one of a plurality of second processing units in said second processing node by a respective one of a plurality of point-to-point links, such that;

a first processing unit in said first processing node is coupled to a fourth processing unit in the second processing node by a first point-to-point link;

a second processing unit in said first processing node is coupled to a fifth processing unit in the second processing node by a second point-to-point link; and

a third processing unit in said first processing node is coupled to a sixth processing unit in the second processing node by a third point-to-point link;

wherein;

operations of said plurality of first processing units include, in order, at least a request phase in which a request is broadcast, a partial response phase in which individual processing units determine their respective responses to said request, and a combined response phase in which a system-wide combined response to said request is distributed; and

a node master processing unit among said plurality of first processing units routes said combined response via each link traversed by said request in a same direction as said request and each node leaf processing unit among said plurality of first processing units routes at least one partial response via a link traversed by said request in an opposite direction to said request.