automatic detection of SPARse matrix computations
in Application programs by pattern MATching techniques

Research project supported by DFG, started in Oct. 1997, ended Sep. 2000.

The SPARAMAT project at the Computer Science Department of the University of Trier (Germany) is researching using run-time and compile-time techniques to recognize sequential matrix operations in C or FORTRAN user code and replace them with faster parallel algorithms (MIMD). To do this, we need sample programs so we can develop an idea what patterns to recognize in user code.

SPARAMAT is the follow-up project to PARAMAT.

Researchers involved:

Prof. Dr. Helmut Seidl
Dr. Christoph Kessler
Craig Smith, M.Sc.