Skip to: Site menu | Main content

Software

To increase the impact of the computational formalisms, algorithms, and codes we develop, we publish their details in journals such as Computer Physics Communications. We will release the associated software under a modified open source license* in which each user is required to cite these publications. A modular software engineering approach will be used in development of the project software. Use of a component standard such as the Common Component Architecture (CCA) will be evaluated for appropriateness for the code developed in this project. The web interface leverages existing grid-portal development frameworks and standards to ensure interoperability with other web services. All software is being managed and maintained using a version control system. Components undergo unit testing for both correctness and performance, and a defect and feature tracking system is be used for managing error reporting and maintenance of all released software. Milestones: Four codes, with full documentation, publicly available by Y3.

* Our license allow full and open use, modification, and extension of our software products, such as described at: http://opensource.org/licenses/ with the minimal restrictions to protect the jurisdiction's ownership rights.





ManyJobs

ManyJobs is a python based tool for managing ManyTasks on widely distributed supercomputing resources. To date it has been used in production mode on the following machines: Abe@NCSA, Lonestar@TACC, Ranger@TACC, QueenBee@LONI, Kraken@NICS, Steele@RCAC, (eric, louie, poseidon, painter, oliver)@LONI, and the CCS Cluster at Tulane's CCS.





BigJob

A Pilot-Job allows the execution of jobs without the necessity to queue each individual job. A pilot job is started through the regular Grid resource manager, provides a container for many sub-jobs, i.e. applications submit these sub-jobs through the pilot job and not the resource manager. A major advantage of this approach is that the waiting time at the local resource manager, which usually significantly contributes the overall time-to-completion, is avoided. It also provides application-level control of the sub-job execution.
The SAGA BigJob framework is a SAGA-based pilot job implementation. Unlike other common pilot job systems SAGA BigJob (i) natively supports MPI job and (ii) works on a variety of back-end systems, generally reflecting the advantage of using a SAGA-based approach.