Dispatching packets on a global combining network of a parallel computer

• US 7,984,450 B2
• Filed: 11/28/2007
• Issued: 07/19/2011
• Est. Priority Date: 11/28/2007
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Archer, Charles J., Almasi, Gheorghe
• Primary Examiner(s): Ho; Andy
• Application Number:
  US11/946,136
• Publication Number:
  US 20090138892A1
• Time in Patent Office: 1,329 Days
• Field of Search: 719/313, 718/106
• US Class Current: 719/313
• CPC Class Codes: G06F 13/387 for adaptation of different data processing systems to different peripheral devices, e.g. protocol converters for incompatible systems, open system

Abstract

Methods, apparatus, and products are disclosed for dispatching packets on a global combining network of a parallel computer comprising a plurality of nodes connected for data communications using the network capable of performing collective operations and point to point operations that include: receiving, by an origin system messaging module on an origin node from an origin application messaging module on the origin node, a storage identifier and an operation identifier, the storage identifier specifying storage containing an application message for transmission to a target node, and the operation identifier specifying a message passing operation; packetizing, by the origin system messaging module, the application message into network packets for transmission to the target node, each network packet specifying the operation identifier and an operation type for the message passing operation specified by the operation identifier; and transmitting, by the origin system messaging module, the network packets to the target node.

Claims

1. A method for dispatching packets on a global combining network of a parallel computer, the parallel computer comprising a plurality of compute nodes, the compute nodes connected for data communications using the global combining network, the global combining network capable of performing collective operations and point to point operations, the method further comprising:

receiving, by an origin system messaging module on an origin compute node from an origin application messaging module on the origin compute node, a storage identifier and an operation identifier, the storage identifier specifying application storage containing an application message for transmission to at least one target compute node, and the operation identifier specifying a message passing operation to be performed by the origin application messaging module;

packetizing, by the origin system messaging module, the application message into network packets for transmission to the target compute node, each network packet specifying the operation identifier and an operation type for the message passing operation specified by the operation identifier; and

transmitting, by the origin system messaging module, the network packets to the target compute node.

2. The method of claim 1 further comprising:

receiving, by a target system messaging module on the target compute node, one of the network packets;

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet; and

executing, by the target system messaging module, a callback function in dependence upon the operation type and the operation identifier.

3. The method of claim 2 wherein:

the operation identifier is stored in a first portion of a payload for the received network packet; and

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet further comprises retrieving the operation identifier from the first portion of the payload for the received network packet without reading the remaining portion of the payload into computer memory.

4. The method of claim 1 further comprising:

receiving, by the origin system messaging module from the origin application messaging module, a registration request for the message passing operation to be performed by the origin application messaging module; and

providing, by the origin system messaging module to the origin application messaging module in response to receiving the registration request, the operation identifier for the message passing operation.

5. The method of claim 1 wherein the operation identifier is stored at a location in each network packet in dependence upon the operation type.

6. The method of claim 1 wherein the operation type specifies whether the message passing operation is a collective operation or a point to point operation.

7. A parallel computer capable of dispatching packets on a global combining network of a parallel computer, the parallel computer comprising an origin compute node and at least one target compute node, the compute nodes connected for data communications using the global combining network, the global combining network capable of performing collective operations and point to point operations, the origin compute node comprising a origin computer processor and origin computer memory operatively coupled to the origin computer processor, the origin computer memory having disposed within it computer program instructions capable of:

receiving, by an origin system messaging module on the origin compute node from an origin application messaging module on the origin compute node, a storage identifier and an operation identifier, the storage identifier specifying application storage containing an application message for transmission to at least one target compute node, and the operation identifier specifying a message passing operation to be performed by the origin application messaging module;

packetizing, by the origin system messaging module, the application message into network packets for transmission to the target compute node, each network packet specifying the operation identifier and an operation type for the message passing operation specified by the operation identifier; and

transmitting, by the origin system messaging module, the network packets to the target compute node.

8. The parallel computer of claim 7 wherein the target compute node comprises a target computer processor and target computer memory operatively coupled to the target computer processor, the target computer memory having disposed within it computer program instructions capable of:

receiving, by a target system messaging module on the target compute node, one of the network packets;

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet; and

executing, by the target system messaging module, a callback function in dependence upon the operation type and the operation identifier.

9. The parallel computer of claim 8 wherein the operation identifier is stored in a first portion of a payload for the received network packet;

and

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet further comprises retrieving the operation identifier from the first portion of the payload for the received network packet without reading the remaining portion of the payload into computer memory.

10. The parallel computer of claim 7 wherein the origin computer memory has disposed within it computer program instructions capable of:

receiving, by the origin system messaging module from the origin application messaging module, a registration request for the message passing operation to be performed by the origin application messaging module; and

providing, by the origin system messaging module to the origin application messaging module in response to receiving the registration request, the operation identifier for the message passing operation.

11. The parallel computer of claim 7 wherein the operation identifier is stored at a location in each network packet in dependence upon the operation type.

12. The parallel computer of claim 7 wherein the operation type specifies whether the message passing operation is a collective operation or a point to point operation.

13. A computer program product for dispatching packets on a global combining network of a parallel computer, the parallel computer comprising a plurality of compute nodes, the compute nodes connected for data communications using the global combining network, the global combining network capable of performing collective operations and point to point operations, the computer program product disposed upon a recordable computer readable medium, the computer program product comprising computer program instructions capable of:

receiving, by an origin system messaging module on an origin compute node from an origin application messaging module on the origin compute node, a storage identifier and an operation identifier, the storage identifier specifying application storage containing an application message for transmission to at least one target compute node, and the operation identifier specifying a message passing operation to be performed by the origin application messaging module;

packetizing, by the origin system messaging module, the application message into network packets for transmission to the target compute node, each network packet specifying the operation identifier and an operation type for the message passing operation specified by the operation identifier; and

transmitting, by the origin system messaging module, the network packets to the target compute node.

14. The computer program product of claim 13 further comprising computer program instructions capable of:

receiving, by a target system messaging module on the target compute node, one of the network packets;

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet; and

executing, by the target system messaging module, a callback function in dependence upon the operation type and the operation identifier.

15. The computer program product of claim 14 wherein

the operation identifier is stored in a first portion of a payload for the received network packet;

and

identifying, by the target system messaging module, the operation type and the operation identifier for the received network packet further comprises retrieving the operation identifier from the first portion of the payload for the received network packet without reading the remaining portion of the payload into computer memory.

16. The computer program product of claim 13 further comprising computer program instructions capable of:

receiving, by the origin system messaging module from the origin application messaging module, a registration request for the message passing operation to be performed by the origin application messaging module; and

providing, by the origin system messaging module to the origin application messaging module in response to receiving the registration request, the operation identifier for the message passing operation.

17. The computer program product of claim 13 wherein the operation identifier is stored at a location in each network packet in dependence upon the operation type.

18. The computer program product of claim 13 wherein the operation type specifies whether the message passing operation is a collective operation or a point to point operation.