Executing a distributed java application on a plurality of compute nodes in accordance with a just-in-time compilation history

Patent Number: 8595742, issued on 2013/11/26    
Applied on 2012/04/12, 13/443,606
Inventor(s): Eric Barsness, David Darrington, Amanda Randles, John Santosuosso
Assignee:

Abstract: Methods, systems, and products are disclosed for executing a distributed Java application on a plurality of compute nodes. The Java application includes a plurality of jobs distributed among the plurality of compute nodes. The plurality of compute nodes are connected together for data communications through a data communication network. Each of the plurality of compute nodes has installed upon it a Java Virtual Machine (`JVM`) capable of supporting at least one job of the Java application. Executing a distributed Java application on a plurality of compute nodes includes: tracking, by an application manager, a just-in-time (`JIT`) compilation history for the JVMs installed on the plurality of compute nodes; and configuring, by the application manager, the plurality of jobs for execution on the plurality of compute nodes in dependence upon the JIT compilation history for the JVMs installed on the plurality of compute nodes.

Claims: 1. A method of executing a distributed software application on a plurality of compute nodes, the software application comprising a plurality of jobs distributed among the plurality of compute nodes, the plurality of compute nodes connected together for data communications through a data communication network, each of the plurality of compute nodes having installed upon it a Virtual Machine (`VM`) capable of supporting at least one job of the software application, the method comprising: tracking, by an application manager, a just-in-time (`JIT`) compilation history for the VMs installed on the plurality of compute nodes, wherein the JIT compilation history includes JIT levels for portions of the software application previously processed on the plurality of compute nodes; and configuring, by the application manager, the plurality of jobs for execution on the plurality of compute nodes in dependence upon the JIT compilation history for the VMs installed on the plurality of compute nodes including: performing for at least one job of the software application: selecting, for that job, one of the compute nodes on which to execute that job in dependence upon the JIT compilation history for the VMs and at least one desired JIT level for that job; and configuring that job on the selected compute node for execution. 2. The method of claim 1 wherein the JIT compilation history further comprises JIT levels for software constructs processed on the plurality of compute nodes. 3. The method of claim 1 wherein the JIT compilation history further comprises JIT levels for software call sequences for particular software methods processed on the plurality of compute nodes. 4. The method of claim 1 further comprising: recording, by each VM supporting at least one job of the software application on the plurality of compute nodes, the JIT compilation history for the jobs supported by that VM; and providing, by each VM supporting at least one job of the software application on the plurality of compute nodes, the recorded JIT compilation history to the application manager. 5. The method of claim 1 wherein executing a distributed software application on a plurality of compute nodes further comprises executing a distributed software application on a parallel computer, the parallel computer comprising the plurality of compute nodes and a service compute node, the application manager installed upon the service compute node, the plurality of compute nodes connected for data communications through a plurality of data communications networks, at least one data communications network optimized for collective operations, and at least one other data communications network optimized for point to point operations. 6. A computing system for executing a distributed software application, the computing system comprising a plurality of compute nodes, the software application comprising a plurality of jobs distributed among the plurality of compute nodes, the plurality of compute nodes connected together for data communications through a data communication network, each of the plurality of compute nodes having installed upon it a Virtual Machine (`VM`) capable of supporting at least one job of the software application, the computing system comprising one or more computer processors and computer memory operatively coupled to the computer processors, the computer memory having disposed within it computer program instructions that when executed by the one or more computer processors cause the computing system to carry out the steps of: tracking, by an application manager, a just-in-time (`JIT`) compilation history for the VMs installed on the plurality of compute nodes, wherein the JIT compilation history includes JIT levels for portions of the software application previously processed on the plurality of compute nodes; and configuring, by the application manager, the plurality of jobs for execution on the plurality of compute nodes in dependence upon the JIT compilation history for the VMs installed on the plurality of compute nodes including: performing for at least one job of the software application: selecting, for that job, one of the compute nodes on which to execute that job in dependence upon the JIT compilation history for the VMs and at least one desired JIT level for that job; and configuring that job on the selected compute node for execution. 7. The computing system of claim 6 wherein the JIT compilation history further comprises JIT levels for software constructs processed on the plurality of compute nodes. 8. The computing system of claim 6 wherein the JIT compilation history further comprises JIT levels for software call sequences for particular software methods processed on the plurality of compute nodes. 9. The computing system of claim 6 wherein the computer memory has disposed within it computer program instructions that when executed by the one or more computer processors cause the computing system to carry out the steps of: recording, by each VM supporting at least one job of the software application on the plurality of compute nodes, the JIT compilation history for the jobs supported by that VM; and providing, by each VM supporting at least one job of the software application on the plurality of compute nodes, the recorded JIT compilation history to the application manager. 10. The computing system of claim 6 wherein executing a distributed software application on a plurality of compute nodes further comprises executing a distributed software application on a parallel computer, the parallel computer comprising the plurality of compute nodes and a service compute node, the application manager installed upon the service compute node, the plurality of compute nodes connected for data communications through a plurality of data communications networks, at least one data communications network optimized for collective operations, and at least one other data communications network optimized for point to point operations. 11. A computer program product for executing a distributed software application on a plurality of compute nodes, the software application comprising a plurality of jobs distributed among the plurality of compute nodes, the plurality of compute nodes connected together for data communications through a data communication network, each of the plurality of compute nodes having installed upon it a Virtual Machine (`VM`) capable of supporting at least one job of the software application, the computer program product disposed upon a computer readable medium, the computer readable medium not comprising a signal, the computer program product comprising computer program instructions that when executed by a computer cause the computer to carry out the steps of: tracking, by an application manager, a just-in-time (`JIT`) compilation history for the VMs installed on the plurality of compute nodes, wherein the JIT compilation history includes JIT levels for portions of the software application previously processed on the plurality of compute nodes; and configuring, by the application manager, the plurality of jobs for execution on the plurality of compute nodes in dependence upon the JIT compilation history for the VMs installed on the plurality of compute nodes including: performing for at least one job of the software application: selecting, for that job, one of the compute nodes on which to execute that job in dependence upon the JIT compilation history for the VMs and at least one desired JIT level for that job; and configuring that job on the selected compute node for execution. 12. The computer program product of claim 11 wherein the JIT compilation history further comprises JIT levels for software constructs processed on the plurality of compute nodes. 13. The computer program product of claim 11 wherein the JIT compilation history further comprises JIT levels for software call sequences for particular software methods processed on the plurality of compute nodes. 14. The computer program product of claim 11 further comprising computer program instructions that when executed by the computer cause the computer to carry out the steps of: recording, by each VM supporting at least one job of the software application on the plurality of compute nodes, the JIT compilation history for the jobs supported by that VM; and providing, by each VM supporting at least one job of the software application on the plurality of compute nodes, the recorded JIT compilation history to the application manager. 15. The computer program product of claim 11 wherein executing a distributed software application on a plurality of compute nodes further comprises executing a distributed software application on a parallel computer, the parallel computer comprising the plurality of compute nodes and a service compute node, the application manager installed upon the service compute node, the plurality of compute nodes connected for data communications through a plurality of data communications networks, at least one data communications network optimized for collective operations, and at least one other data communications network optimized for point to point operations.