One more project I did in UPC for the course of parallel programming models and algorithms was to implement two parallel versions of the Smith-Waterman sequence alignment algorithm and test them on an ALTIX machine. The different implementations used parallelization techniques such as blocking and interleaving. We finished up the report with the performance model of each implementation for two different network topologies.
In the end, I got more comfortable with performance models and network topologies. We would have liked to finish it with a more detailed conclusion but the waiting times to get it running on the ALTIX machines dictated the time with had to write it.
So here is our report :