Job Description

Mary is responsible for the management and development of the support infrastructure for research faculty, research staff, and graduate students in their use of high performance computing resources at Georgia Tech. She supervises a small staff of two research scientists and several graduate students. In addition to performing the administrative duties for the group, she acts as a user consultant.

Her administrative duties include coordinating the support of a wide variety of users and applications on a number of high performance systems including vector, parallel, and superscalar architectures. This includes personnel management and guidance, budget preparation and management, policy and procedure development and control, approval and creation of new accounts, monitoring current system usage, and preparation of mangement reports. She also works with faculty groups from across the campus in developing strategies for future directions in high performance computing as well as policies for existing platforms. Mary is also the Georgia Tech Internet2 Application Liason,,

Her general user support duties include the advisement of users in determining the appropriate systems and tools for his or her research with follow-up assistance in applying for time if necessary; development and/or presentation of training material on topics related to vector and parallel computer architectures; acquisition and/or preparation of publications such as User's Guides, tutorials, and reference handbooks on general topics such as code optimization and library usage as well as advanced topics such as data decomposition and techniques for algorithm restructuring; consultation with users on programming problems, general usage, application ports.

Mary also participates in sponsored research projects. In order to maintain skills on the use of high performance systems it is beneficial to participate in such efforts. This also allows the staff to offer their expertise to various research groups who may not have certain skills on high performance computing systems that are required in a solicitation.

Research Interests

Earth and Atmospheric Sciences. While working in high performance computing support, I have been exposed to a vast number of scientific and engineering research projects. This has been one of the most interesting and fulfilling aspects of my career. I have worked on the parallelization of a global atmospheric model of the transport of chemical constituents in the stratosphere.

Parallel Processing Techniques for Research Projects in Science and Engineering. This includes optimizing application programs which use large amounts of data or need massive amounts of cpu and memory to calculate results of any significance. This means tailoring the program code to fit the machine architecture as closely as possible.

Computer Architectures and Performance. This involves keeping current on the newest technologies available, the way in which these technologies are used to implement new computer systems, and what types of applications are best suited to which types of new architectures.

Video Applications over Advanced Networks. Video Conferencing and Video on Demand for meetings, collaboration, and distance learning will become a reality with the new networks under development today. I am working on two projects which involve video over advanced networks:

  • Video Development Initiative (ViDe) is a multi-institutional research group whose primary goal is to work with vendors to ensure standards based video clients for conferencing and archiving over IP in an academic environment.
  • SEPSCORE Bandwidth Testing Project brings together nine sites whose goal is to deliver training modules on nine different high performance computing systems over the advanced networks. The modules will contain features such as video on demand, video conference consultants, animations to enhance content delivery, chats and bulletin boards, etc.

Network Access to Computational and Data Resources. The largest high performance computing systems, data repositories, and other unique resources are generally clustered in various national labs, including NSF, DOE, DOD, and NASA. Distributed parallel computing systems are also becoming widely used. But this type of arrangement is only as good as the scientist's or engineer's access to the resources. Therefore the networking between an institution and the research laboratory or between laboratories themselves is extremely important and therefore Mary is working on the Internet 2 project.

Clustering of Heterogeneous Computers and Superworkstations for Compute Intensive Applications. The advent of powerful workstations and small parallel systems at costs that are affordable by many institutions and laboratories has brought about the idea of clustering. Clustering means tying a number of these small, yet powerful systems together in order to work on a large application. This type of solution for running large applications, though having the added complexity of workload scheduling and networking, shows the promise of considerable computer power in a heterogeneous processor environment.

Contact Information

Mary Trauner
Office of Information Technology
Atlanta, GA 30332-0710

Rich Building, Room 326

(404)894-6166 (office)
(404)894-4690 (fax)

mary.trauner@oit.gatech.edu

Other Interests

In the time that's left, Mary enjoys hiking, tall places, mountain biking, mountaineering, mushing, whitewater canoing, wildlife viewing, blues music, and various forms of country (folk) dancing (clogging, contra, zydeco, and cajun.)