FCT Programs
About The Program Governance Media Contacts
Friday, May 16, 2008
Digital Media Advanced Computing Mathematics UTEN
Research Agenda for Further Interactions

The following areas will be jointly explored in order to find specific research topics of common interest:

  • Computational Science: modeling, simulation, optimization, performance analysis and tuning, numerical methods for parallel and distributed computing;
  • Computer System: paradigms and algorithms, system performance analysis, middleware, tools and technologies, and development of useful environments to the user.
  • Applications: in science, engineering, economics, civil protection and risk management;

In addition, TACC and the Portuguese institutions will work together on deployment and operational issues for advanced computing, such as software, tools evaluation and secure access. All projects in these areas will include education, training, outreach activities and will consider opportunities for technology transfer, as Figure 2 shows.

In the area of Computational Science, work will be developed closely with teams of computational scientists and applied mathematicians from both UT-Austin and Portugal. Joint work will build on previous experience with interdisciplinary research at all the Portuguese universities involved in the partnership and will focus on the development of parallel numerical methods for science and engineering applications.

In the area of Computer Science, the main research issues to pursue include models and tools for programming, compilation, execution and debugging, web services, workflows, (utility driven) resource allocation in networks and global properties of large computer networks. Issues in fault-tolerance, like autonomic computing, self-healing, software aging, rejuvenation, dependability, and quality of service in advanced computing, will also be addressed. Overlay and peer-to-peer networks will also be studied as well as broadband wireless technology for advanced computing.

The research agenda will also include work on load balancing and resource allocation across a shared computing and communication infrastructure. For example, there is a need to extend traditional quality of service (QoS) metrics that have primarily been defined for communications services, so as to enhance the performance of an advanced and grid computing application, or multiple applications that share computing/communication infrastructure.

Research towards formulating and extending utility-based resource allocation for dynamic networks, and in particular for metrics that are relevant to the applications at hand, e.g., delays in transferring large amounts of raw data to remote sites, or approaches to routing and network caching of shared data among distributed sites will be pursued.

In the area of Deployment and Operations, collaboration will be aimed at improving cluster and grid computing environments by exchanging information about approaches and technologies, by developing best practices, and by investigating issues of U.S.-Europe advanced interoperability. The collaboration will also provide guidelines for those performing R&D using clusters and grids so that technologies are usable and provide the needed functionality.

In the area of Applications, the collaboration will consider the following:

  • Study the relationship between advanced computing and power systems, and how the former can be used to enhance the performance of the latter;
  • Optimize models and simulations using the advantages offered by advanced and grid computing technology (in fields such as oceanography, new materials, nanotechnology, flooding, and forest fires);
  • Develop distributed data warehouse for simulation data obtained from protein folding and unfolding;
  • Develop algorithms for large-scale parallelization of image processing and visualization (for medical applications, cultural heritage);
  • Solve challenges posed by the analysis of terabytes of raw data (e.g., telescope information data processing about stars);
  • Apply advanced and grid computing to research in environmental sciences, material sciences and geological sciences;
  • Develop software verification and validation methods for parallel algorithms and result analysis.
  • Develop web portals for user-friendly access to grids and scientific results;
  • Test existing parallel codes using different advanced environments;

All research programs will include faculty and student exchange, co-supervision of Ph.D. students, and hosting of exchange post-doc fellows. An annual workshop in the field of advanced computing will be co-organized between UT-Austin and Portugal in the second and following years, which should include a Ph.D. doctoral consortium. A series of regular seminars of visiting faculty and over the Internet on advanced computing between Portugal and UT-Austin is also intended.

09.Oct.2008
03.Oct.2008
21.Jul.2008
15.Jul.2008
14.Jul.2008
10.Jul.2008
16.Jun.2008
16.Jun.2008
26.May.2008
03.Apr.2008
14.Mar.2008