CLUSTER COMPUTING SUPPORT FOR APPLICATION PROGRAMS

• US 20070288935A1
• Filed: 05/04/2007
• Published: 12/13/2007
• Est. Priority Date: 06/13/2006
• Status: Active Grant

• Current Assignee: Advanced Cluster Systems, Inc.
• Original Assignee: Advanced Cluster Systems, Inc.
• Inventors: Tannenbaum, Zvi, Dauger, Dean
• Granted Patent: US 8,082,289 B2
• Time in Patent Office: Days
• US Class Current: 719/319
• CPC Class Codes: G06F 15/76 Architectures of general purpose stored program computers with program plugboard G06F15/08; multicomputers G06F15/16
• CPC Class Codes: G06F 9/5072 Grid computing
• CPC Class Codes: G06F 9/54 Interprogram communication
• CPC Class Codes: H04L 41/04 Network management architectures or arrangements

Abstract

A computer cluster system comprising a plurality of nodes and a software package comprising a user interface and a kernel for interpreting program code instructions is provided. In one embodiment, a cluster node module is configured to communicate with the kernel and other cluster node modules. The cluster node module accepts instructions from the user interface and interprets at least some of the instructions such that several cluster node modules in communication with one another and with a kernel can act as a computer cluster.

Claims

1. A computer cluster comprising:

a first processor;

a second processor;

a third processor;

at least one computer-readable medium in communication at least one of the first processor, the second processor, or the third processor;

a first kernel residing in the at least one computer-readable medium, said first kernel configured to translate commands into code for execution on the first processor;

a first cluster node module residing in the at least one computer-readable medium, said first cluster node module configured to send commands to the first kernel and receives commands from a user interface;

a second kernel residing in the at least one computer-readable medium, said second kernel configured to translate commands into code for execution on the second processor;

a second cluster node module residing in the at least one computer-readable medium, said second cluster node module configured to send commands to the second kernel and communicates with the first cluster node module;

a third kernel residing in the at least one computer-readable medium, said third kernel configured to translate commands into code for execution on the third processor; and

a third cluster node module residing in the at least one computer-readable medium, said third cluster node module configured to send commands to the third kernel and configured to communicate with the first cluster node module and the second cluster node module;

wherein the first cluster node module comprises a data structure in which messages originating from the second and third cluster node modules are stored.

2. The computer cluster of claim 1, wherein each of the first kernel, the second kernel, and the third kernel comprises a Mathematica kernel.

3. The computer cluster of claim 1, wherein each of the first cluster node module, the second cluster node module, and the third cluster node module comprises program code for implementing automatic load balancing on the first processor, the second processor, and the third processor.

4. The computer cluster of claim 1, further comprising a communications network for connecting at least two of the first processor, the second processor, or the third processor to one another.

5. The computer cluster of claim 1, wherein the at least one computer-readable medium comprises a first computer-readable medium coupled to at least the first processor and a second computer-readable medium coupled to at least the second processor, wherein the first kernel and the first cluster node module reside in the first computer-readable medium, and wherein the second kernel and the second cluster node module reside in the second computer-readable medium.

6. The computer cluster of claim 1, further comprising a cluster configuration module that initializes at least one of the first cluster node module, the second cluster node module, or the third cluster node module.

7. The computer cluster of claim 1, wherein the second cluster node module communicates with the first cluster node module using commands that implement at least a portion of a Message-Passing Interface.

8. The computer cluster of claim 1, wherein the data structure of the first cluster node module comprises a received message queue.

9. The computer cluster of claim 8, wherein the first cluster node module further comprises a message receiving queue, in which data specifying the location to which the first cluster node module can expect to receive an expression and an identifier of a processor from which the expression is expected to be sent are stored.

10. The computer cluster of claim 1, wherein the second cluster node module further comprises a received message queue in which messages originating from the first and third cluster node modules are stored.

11. The computer cluster of claim 1, wherein the first processor and the second processor reside in the same computer system.

12. The computer cluster of claim 1, wherein the first processor and the second processor reside on the same die.

13. The computer cluster of claim 1, wherein the first cluster node module comprises an advanced functions module, and wherein the advanced functions module comprises a call that evaluates an expression across the computer cluster in parallel.

14. The computer cluster of claim 13, wherein the advanced functions module comprises a call that calculates a Fourier transform across the computer cluster in parallel.

15. The computer cluster of claim 13, wherein the advanced functions module comprises one or more calls that implement guard-cell operations.

16. The computer cluster of claim 13, wherein the advanced functions module comprises one or more calls that evaluate matrix operations across the computer cluster in parallel.

17. A computer cluster comprising:

a plurality of nodes, wherein each node is configured to access a computer-readable medium comprising program code for a user interface and program code for a single-node kernel module configured to interpret user instructions;

a plurality of cluster node modules, wherein each cluster node module is configured to communicate with a single-node kernel and with one or more other cluster node modules, to accept instructions from the user interface, and to interpret at least some of the user instructions such that the plurality of cluster node modules communicate with one another in order to act as a cluster; and

a communications network to connect the nodes;

wherein one of the plurality of cluster node modules returns a result to the user interface.

18. The computer cluster of claim 17, wherein at least one of the plurality of cluster node modules comprises program code for storing data to one or more storage disks.

19. The computer cluster of claim 18, wherein the at least one of the plurality of cluster node modules further comprises program code for retrieving stored data from one or more storage disks.

20. The computer cluster of claim 17, wherein at least one of the plurality of cluster node modules comprises program code for collecting data from the plurality of cluster node modules and writing the data to a storage device associated with a single computer system.

21. The computer cluster of claim 17, wherein at least one of the plurality of cluster node modules comprises program code for reading data from a storage device associated with a single computer system and distributing the data to the plurality of cluster node modules within the computer cluster.

22. The computer cluster of claim 17, wherein at least one of the plurality of cluster node modules comprises program code for writing temporary data to a storage device associated with the same computer system on which the respective cluster node module is running.

23. The computer cluster of claim 17, wherein the communications network comprises at least one of a local bus, a local-area network, a wide-area network, an Ethernet network, a Token Ring network, a Frame Relay network, a Fiber Distributed Datalink Interface network, or an Asynchronous Transfer Mode network.

24. The computer cluster of claim 17, wherein the user interface comprises a Mathematica front end.

25. The computer cluster of claim 17, wherein the user interface comprises a command line.

26. The computer cluster of claim 17, wherein each of the plurality of cluster node modules comprises a toolkit comprising library calls that implement at least a portion of a Message-Passing Interface.

27. The computer cluster of claim 17, wherein each of the plurality of cluster node modules comprises a toolkit comprising one or more high-level cluster computing commands.

28. The computer cluster of claim 17, wherein the cluster computer system comprises at least one of a plurality of Macintosh®

computers, a plurality of Windows®

-based computers, or a plurality of computers running a Unix or Linux operating system.

29. A method of evaluating a command on a computer cluster comprising:

communicating a command from at least one of a user interface or a script to one or more cluster node modules within the computer cluster;

for each of the one or more cluster node modules, communicating a message based on the command to a respective kernel module associated with the cluster node module;

for each of the one or more cluster node modules, receiving a result from the respective kernel module associated with the cluster node module; and

for at least one of the one or more cluster node modules, responding to messages from other cluster node modules.

30. The method of claim 29, wherein communicating a message based on the command to a respective kernel module associated with the cluster node module comprises communicating a specially identified message to the respective kernel module.

31. The method of claim 29, further comprising, for at least one of the cluster node modules, forwarding the result to at least one of a user interface or a script running on a same computer system as the cluster node module.

32. The method of claim 29, further comprising, for at least one of the cluster node modules, communicating the result in a message to one or more other cluster node modules.

33. The method of claim 29, wherein responding to messages from other cluster node modules comprises:

forwarding messages from the other cluster node modules to a received message queue;

matching data from each entry in a message receiving queue with entries in the received message queue;

combining data from the message receiving queue with matching data in the received message queue; and

marking the matching data as completed.

34. The method of claim 29, wherein communicating a command from at least one of a user interface or a script to one or more cluster node modules within the computer cluster comprises communicating an instruction from a Mathematica Front End to a first cluster node module, wherein the first cluster node module forwards the instruction to other cluster node modules running on the computer cluster.

35. The method of claim 34, wherein the first cluster node module forwards the instruction to other cluster node modules running on the computer cluster using a command from a Message-Passing Interface.

36. The method of claim 29, wherein each of the one or more cluster node modules communicates with a respective kernel module using MathLink.

37. A computing system for executing Mathematica code on multiple nodes comprising:

a first node module in communication with a first Mathematica kernel executing on a first node;

a second node module in communication with a second Mathematica kernel executing on a second node; and

a third node module in communication with a third Mathematica kernel executing on a third node;

wherein the first node module, the second node module, and the third node module are configured to communicate with one another using a peer-to-peer architecture.

38. The computing system of claim 37, wherein each of the first node module, the second node module, and the third node module comprises:

a data structure for maintaining messages originating from other node modules; and

a data structure for maintaining data specifying;

a location to which a message is expected to be received; and

an identifier for a node from which the message is expected to be sent.