HPC-GAP: High Performance Computational Algebra and
HPC-GAP is the EPSRC funded project to reengineer the software for
computation in algebra and discrete mathematics to take advantage of
the power of current and future high-performance computers. Our main
focus is on the GAP system
and the more recent SymGridPar
middleware, which provide flexible and effective computation on
single processors and small clusters. We will adapt the software to
efficiently use large clusters of multi-core processors to perform
larger computations. To demonstrate the effectiveness of our
adaptations we will apply our new software to problems from a number
of important areas of pure mathematics.
The project was started in September 2009 and will run for four years.
- Reimer Behrends, Alexander Konovalov, Steve Linton, Frank Lübeck, Max Neunhöffer. Parallelising the computational algebra system GAP. PASCO 2010: 177-178.
- Reimer Behrends, Alexander Konovalov, Steve Linton, Frank Lübeck, Max Neunhöffer. Towards High-Performance Computational Algebra with GAP. ICMS 2010: 58-61.
- The HPC-GAP project. Parallel programming support in GAP. ACM Communications in Computer Algebra, Vol.46, No.4, Issue 182, December 2012. PDF
- Chris Maynard. Comparing UPC and one-sided MPI: A distributed hash table for GAP. PDF
- Reimer Behrends, Parallelising the computational algebra system GAP. PASCO 2010, Grenoble, France, July 2010 PDF Slides
- Steve Linton, ICMS 2010, Kobe, Japan, September 2010 PDF Slides
- Steve Linton, Computational Group Theory Workshop, Oberwolfach, Germany, August 2011 PDF Slides