Faster Smith-Waterman database searches with inter-sequence SIMD parallelisation
2011

Faster Smith-Waterman Database Searches with SIMD Parallelisation

publication Evidence: high

Author Information

Author(s): Torbjørn Rognes

Primary Institution: Department of Informatics, University of Oslo

Hypothesis

Can further speed be gained in the Smith-Waterman algorithm through new parallelisation approaches?

Conclusion

The new tool SWIPE allows Smith-Waterman database searches to run more than six times faster than previous implementations.

Supporting Evidence

  • SWIPE achieved a speed of 106 billion cell updates per second on a dual Intel Xeon X5650 six-core processor system.
  • SWIPE was about six times faster than Farrar's previous implementation.
  • SWIPE was about twice as fast as BLAST using the BLOSUM50 score matrix.

Takeaway

This study created a new tool that helps compare DNA sequences much faster than before, making it easier to find similarities.

Methodology

The study implemented a new algorithm called SWIPE that processes residues from multiple database sequences in parallel using SIMD technology.

Limitations

The current version of SWIPE cannot search databases larger than about 4 billion amino acids.

Digital Object Identifier (DOI)

10.1186/1471-2105-12-221

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication