Design and study of a massively multi threaded shared memory architecture
Universite de Rennes
Internship report, Research master’s degree in computer science, Universite de Rennes 1, 2011
Most biocomputing problems require a high processing power with high memory needs while showing massive parallelism opportunities. Unfortunately, although advances are made in software parallelism, current architectures do not provide a transparent way to use this parallelism at its full potential. We thus started to design an massively parallel megathreaded architecture that would match biocomputing algorithms requirements both in terms of processing power and memory.
December 12, 2011 by hgpu
