Cell-SWat: Modeling and Scheduling Wavefront Computations on the Cell Broadband Engine

Aji, Ashwin and Feng, Wu-chun and Blagojevic, Filip and Nikolopoulos, Dimitrios (2008) Cell-SWat: Modeling and Scheduling Wavefront Computations on the Cell Broadband Engine . In Proceedings Computing Frontiers.

Abstract

This paper contributes and evaluates a model and a methodology for implementing parallel wavefront algorithms on the Cell Broadband Engine. Wavefront algorithms are vital in several application areas such as computational biology, particle physics, and systems of linear equations. The model uses blocked data decomposition with pipelined execution of blocks across the synergistic processing elements (SPEs) of the Cell. To evaluate the model, we implement the Smith-Waterman pairwise sequence alignment algorithm as a wavefront algorithm and present key optimization techniques that significantly enhance the vector processing capabilities of the SPE. Our results show perfect linear speedup for up to 16 SPEs on the QS20 dual-Cell blades, and our model is highly scalable for more cores, if available. The accuracy of our model is within 3% of the measured values on average. We then test our model in a throughput-oriented experimental setting, where we couple our model with scheduling techniques that exploit parallelism across the simultaneous execution of multiple sequence alignments. Using our model, we are able to improve the throughput of realistic multi-sequence alignment workloads by up to 8% compared to FCFS (first-come, first-serve), by trading off parallelism within sequence alignments with parallelism across sequence alignments.