This work investigates a range of techniques for automatically mapping CSP process networks onto an MPI cluster. A CSP library was developed to provide the necessary functionality for implementing CSP-based concurrent applications on top of MPI. The library enables seamless communication between processes on the same node and processes across different nodes. A new configuration language was implemented to provide a straightforward way to map processes onto cluster nodes. This was designed in such a way to allow for mapping the same application using different mapping algorithms without having to recompile the application. The resulting proof-of-concept system was then used to evaluate the suitability of well-known graph partitioning algorithms for distributing a suite of CSP-based applications across a compute cluster, with the aim of reducing application execution time in each case. The experimental results are presented in summary form and briefly analyzed.
IOS Press, Inc.
6751 Tepper Drive
Clifton, VA 20124
Tel.: +1 703 830 6300
Fax: +1 703 830 2300 firstname.lastname@example.org
(Corporate matters and books only) IOS Press c/o Accucoms US, Inc.
For North America Sales and Customer Service
West Point Commons
Lansdale PA 19446
Tel.: +1 866 855 8967
Fax: +1 215 660 5042 email@example.com