S. I. Hariharan and Gerald Young
Department of Mathematics
 

Funding Sources
Description of the Project
High Performance Computing Source
Projected Benefits of the Project
Collaborations
Experiences and Outcomes
 

Funding Sources

NASA Grant NCC3-494, NSFDMS-953201 (Gerald Young), Ohio Super Center, NASA (Cray).

[top]
 

Description of the Project

This project involves the development of interfacial motions in materials problems.  Specifically phase transition problems are of key interest.  We are comparing sharp interface models with phase field models in certain asymptotic parameter regimes which can be analyzed without large-scale computations.  Outside these regimes large-scale computations will be needed.  The analysis of development of solidifying fronts as applied to thin films formation have received considerable attention in recent times.  In our earlier work, we investigated a practical solution of chemical vapor deposition (CVD) in conjunction with a vertical CVD reactor.  Other ongoing related work are found in references.

In addition, full three-dimensional time dependent flow models in open domains are also being investigated.  These include capturing nonlinear effects such as shocks in compressible flows.

[top]
 

High Performance Computing Source

This project requires high speed network access for computations, file transfers and a great deal of interactive compiling on remote supercomputers.  Our connection is to the Ohio Super Computer Center and NASA (Cray).

[top]
 

Projected Benefits of the Project

Projected benefits include determination of choices of these models for a given processing configuration.  Currently, data files that are transferred between NASA Lewis Research Center and the Applied Mathematics Computation lab can be of the order 2-4 Gigabytes depending on the model.  These large sets of data are being transported via tapes.

[top]
 

Collaborations

We have ongoing collaborations with the University of New Mexico, University of Notre Dame, Northwestern University, Boston University, and NASA Lewis Research Center.  These collaborations often require high speed network access for computations as well as for file transfers.

[top]
 

Experiences and Outcomes

Hariharan performed a test of I2 speed within and off the campus on Saturday February 12 at 5:49 P.M..  He sent a 1.7 GB file from Ayer hall to Engineering in 10 minutes.  This indicates that the package was moving at 23 Mb/s, but 622 Mb/s was expected speed.  He sent the same file to NASA-Glenn in two hours and 10 minutes.  We concluded that this package did not go over I2 because NASA-Glenn does not have an I2 connection.

This test told us two things:

[top]