Systematic Mapping of Reed–Solomon Erasure Codes on Heterogeneous Multicore Architectures

The use of erasure codes could radically improve the availability of distributed storage in comparison to replication systems with similar storage and bandwidth requirements. The investigation reported in this chapter confirms the advantage of using modern heterogeneous multi/many‐core architectures, especially GPUs, for the efficient implementation of the Reed‐Solomon erasure codes. The chapter investigates how to map systematically the RS erasure codes onto heterogeneous multicore architectures. It starts with the Cell/B.E. architecture‐an innovative solution which is significantly different from the conventional multicore architectures. The chapter focuses on extending the methods developed for the cell multicore on heterogeneous computing platforms combining general purpose multicore CPUs and modern GPU accelerators. The resulting methods allow adaptation of the underlying numerical algorithms to internal characteristics of these platforms. The investigation carried out in the chapter confirmed the advantage of using modern multicore architectures for the efficient implementation of the classic RS erasure codes.