Exploring New Search Algorithms and Hardware for Phylogenetics: RAxML Meets the IBM Cell

  • A. Stamatakis
  • F. Blagojevic
  • D. S. Nikolopoulos
  • C. D. Antonopoulos
Article

DOI: 10.1007/s11265-007-0067-4

Cite this article as:
Stamatakis, A., Blagojevic, F., Nikolopoulos, D.S. et al. J VLSI Sign Process Syst Sign Im (2007) 48: 271. doi:10.1007/s11265-007-0067-4

Abstract

Phylogenetic inference is considered to be one of the grand challenges in Bioinformatics due to the immense computational requirements. RAxML is currently among the fastest and most accurate programs for phylogenetic tree inference under the Maximum Likelihood (ML) criterion. First, we introduce new tree search heuristics that accelerate RAxML by a factor of 2.43 while returning equally good trees. The performance of the new search algorithm has been assessed on 18 real-world datasets comprising 148 up to 4,843 DNA sequences. We then present the implementation, optimization, and evaluation of RAxML on the IBM Cell Broadband Engine. We address the problems and provide solutions pertaining to the optimization of floating point code, control flow, communication, and scheduling of multi-level parallelism on the Cell.

Keywords

phylogenetic inference maximum likelihood RAxML IBM cell 

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • A. Stamatakis
    • 1
  • F. Blagojevic
    • 2
  • D. S. Nikolopoulos
    • 2
  • C. D. Antonopoulos
    • 3
  1. 1.School of Computer and Communication SciencesÉcole Polytechnique Fédérale de Lausanne LausanneSwitzerland
  2. 2.Department of Computer Science, Center for High-end Computing SystemsVirginia Tech BlacksburgUSA
  3. 3.Department of Computer and Communications EngineeringUniversity of ThessalyVolosGreece

Personalised recommendations