|Position Title||Petascale Scientific Parallel Software Developer of the Adaptive Simulation Field Library (ASF)|
|Summary||Design and implementation of Adaptive Simulation Field library in parallel using MPI/OpenMP hybrid model with application to multiscale biotissue and energy (plasma modeling) applications.|
|Job Description||The Scientific Computation Research Center is|
implementing a library aiming at:
1) supporting the storage data associated with physical fields (temperature, displacement, kinetic energy, ...) and higher order tensors of various quantities of interest.
2) carrying various operations on the defined fields (interpolation, transformation, ...)
3) supporting general orders of representation from piecewise constant to high-order including higher order continuity
4) parallel execution on exascale machines
The student will participate in the extension of the design and implementation of the Adaptive Simulation Field (ASF) library. This internship will specifically focus on extending the design and
implementation of the library toward parallel computing.
The first part of this work will consist of extending ASF interface with an ASFp parallel interface component. Its implementation will consist of defining a reduced data structure that will only maintain a minimum set of parallel data. The parallel model of implementation to be used for such a new component will be through and MPI/OpenMP hybrid model.
Execution of the library will be carried out on a large number of processors (100k processors) and evaluation of its performance using tuning software will be performed on multiscale biotissue and energy (plasma modeling) applications.
* Design and C++ implementation of ASF library with specific consideration of providing a set of components that can interact with various domain discretizations. Initial emphasis will be on unstructured meshes.
* Design and implementation of tests (unit and stress tests)
* Implementation of use test cases to be executed on Blue Waters and other massively parallel machines.
* Tuning of the library to improve its efficiency
* Working experience in C/C++, MPI, OpenMP, SVN/CVS and Agile method
* Experience in using debugger and tuning software
* Experience in working on supercomputer
* Ability to work in a development team
Ability to work in Troy, NY during the school year to continue working with the Scientific Computation Research Center team of faculty, research staff and students working on interoperable components for massively parallel scientific computation.
|Location||Scientific Computation Research Center (SCOREC)|
Rensselaer Polytechnic Institute (RPI)
Low Center for Industrial Innovation, CII-4011
110 8th Street
Troy, NY 12180-3590