Evaluation and deployment of software tools that support the community at-large helps connect the NCSA CSM Group to a wide variety of people and projects. Opportunities for collaboration present themselves naturally.

Continued studies of new research codes and commercial software products will give utmost consideration to the ability of any tool to accommodate a wide cross-section of users and applications with a minimum of training and effort. We do not have the resources to collaborate with every user of software products we might recommend for deployment.

Design optimization and parallel processing tools have been targeted precisely because interest in this type of software is characteristic of the sort of research groups we want to discover. They are more likely to be close to a breakthrough than others, either in terms of performance or functionality.

PROJECT OBJECTIVES

Two software products have been identified and acquired for in-depth analyses:

• CAPTools, The University of Greenwich (UK) and NASA Ames
• iSIGHT, Engineous Software, Inc.

CAPTools is a data dependency analyzer for serial Fortran programs. It is capable of generating OpenMP and message passing versions of the input source code. iSIGHT is a design optimization and job control framework designed to drive multidisciplinary applications.

Other basic goals of this project include:

• Increased utilization of large multiprocessor computers by deploying effective parallelization tools
• Development of interest in design optimization techniques for achieving common engineering goals such as improved durability or performance
• Construction of a multidisciplinary simulation code that spans MSRC boundaries

Collaboration within ASC and the Army Research Laboratory (ARL) will be developed first with MSRC staff. Integration of parallel applications into the iSIGHT framework will be undertaken in a collaborative manner between the CSM, PTES, and IMT CTAs.

In CY5, the NCSA CSM Group will begin to direct the integration of an optimization model that spans MSRCs. The state of network computing in CY6 will determine if the resultant application reaches production level or remains a proof-of-concept code.

DELIVERABLES

Deliverable components of this project include:

• Installation and maintenance of the CAPTools and iSIGHT packages
• Establishment of an iSIGHT users registry to promote multidisciplinary collaboration
• On-site training, one event per year, each package, PicTel transmission to ARL

While we aim to promote self-sufficiency in the user groups associated with these packages, the need to demonstrate concept codes (develop success stories) and make new contacts is also important to the success of this project. The following deliverable addresses this need:

• Identify a number of new parallel porting efforts and one new optimization project for provision of detailed consulting with the NCSA CSM Group.

All research codes employed during the evaluation, deployment and collaboration phases of this project will be returned to the DoD groups responsible for their development (5 simcodes to date).

CUSTOMERS

Customers and end users for the CAPTools software includes all individuals and engineering groups currently maintaining serial FORTRAN codes. Prospective iSIGHT users include all individuals or engineering groups with recognizable multiphysics interfaces.

End users associated with the evaluation and demonstration codes include

Arnold Mayer, Gregory Czarnecki, and Jeffrey Calcaterra (AFRL/VAVS)
Miguel Visbal, Raymond Gordnier, and Reid Melville (AFRL/VA)
Nicholas Pagano and Gregory Schoeppner (AFRL/MLBM).

BENEFIT TO THE WARFIGHTER

Effective transitioning of serial Fortran codes to the high-performance, multiprocessor computing systems promotes the study of larger and more complicated problems.

Integration of discipline-specific analysis tools into coupled physics models enables simulation of real-world weapons systems with unprecedented levels of accuracy.

PROJECT DEPENDENCIES AND SCOPE

Success will depend on continued support for the University of Greenwich and acquisition of adequate licensing and training arrangements. A modicum of travel has also been budgeted to promote accomplishments associated with the use of these tools.

RISK ELEMENT