Combining execution pipelines to improve parallel implementation of HMMER on FPGA

Abstract : HMMER is a widely used tool in bioinformatic, based on the Profile Hidden Markov Models. The computation kernels of HMMER, namely MSV and P7Viterbi are very compute intensive, and their data dependencies if interpreted naively, lead to a purely sequential execution. In this paper, we propose a original parallelization scheme for HMMER by rewriting the mathematical formulation, to expose hidden potential parallelization opportunities. Our parallelization scheme targets FPGA technology, and our architecture can achieve 10 times speedup compared with the latest HMMER3 SSE version, without compromising on the sensitivity of original algorithm.
Type de document :
Article dans une revue
Microprocessors and Microsystems: Embedded Hardware Design (MICPRO), Elsevier, 2015, 39, pp.457-470. 〈http://www.sciencedirect.com/science/article/pii/S014193311500085X〉. 〈10.1016/j.micpro.2015.06.006〉
Liste complète des métadonnées

Littérature citée [37 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01235328
Contributeur : Dominique Lavenier <>
Soumis le : jeudi 10 décembre 2015 - 08:57:51
Dernière modification le : mercredi 11 avril 2018 - 02:00:44

Identifiants

Citation

Naeem Abbas, Steven Derrien, Sanjay Rajopadhye, Patrice Quinton, Alexandre Cornu, et al.. Combining execution pipelines to improve parallel implementation of HMMER on FPGA. Microprocessors and Microsystems: Embedded Hardware Design (MICPRO), Elsevier, 2015, 39, pp.457-470. 〈http://www.sciencedirect.com/science/article/pii/S014193311500085X〉. 〈10.1016/j.micpro.2015.06.006〉. 〈hal-01235328〉

Partager

Métriques

Consultations de la notice

650