Skip Navigation
Text:
Increase font size
Decrease font size

    U.S. - Germany Cooperative Research Effort: High-Performance Execution of Nested Data Parallelism in Fortran Programs

    Principal Investigators: Jan F. Prins and Siddhartha Chatterjee
    Funding Agency: National Science Foundation
    Agency Number: INT-9726317

    Abstract
    We propose an international collaboration between research groups at UNC-Chapel Hill and the Technische Universitat Berlin in the area of high-performance computing. The aim of this collaboration is to integrate nested parallelism into HPF and Fortran 95. Nested parallelism is an important mechanism for expressing the efficient algorithms needed with increasingly large scientific computations, but is currently not supported in performance-oriented programming languages because the efficient execution of such constructions is problematic. Both research groups have been investigating the technical problems that arise in providing efficient parallel execution of nested data parallel programs, and we believe that with a joint effort the problems can be successfully addressed. The proposed collaboration toward this goal brings together two active research groups with complementary strengths and facilities. Combining the expertise of these two relatively small groups through the exchange of scientists will yield the critical mass needed to achieve our goals.

    Document Actions