[Bioclusters] mpiBLAST 1.2.0
Jason Gans
bioclusters@bioinformatics.org
Tue, 18 Nov 2003 17:46:56 -0500
The mpiBLAST development team is pleased to announce the
availability of mpiBLAST 1.2.0. New features in this release
include:
* Alignments are output for translated searches.
* Incremental database updates.
* Distributed sequence lookup during output formatting.
* Commandline options now match the latest version of NCBI-BLAST.
* Scripts are provided to interface mpiBLAST with the NCBI web front
end via the PBS scheduling system.
mpiBLAST is a parallelization of NCBI BLAST that replaces formatdb and
blastall with versions that use MPI to execute in parallel on a cluster
of computers. There are two primary advantages to using mpiBLAST versus
traditional BLAST. First, mpiBLAST splits the database across each node
in the cluster. Because each node's segment of the database is smaller
it can usually reside in the buffer-cache, yielding a significant
speedup
due to the reduction of costly disk I/O. Second, because interprocessor
communication demands are low, BLAST users can efficiently use low-cost
Beowulf clusters .
Source code for both Unix and Windows is freely available at
http://mpiblast.lanl.gov
Please direct questions and bug reports to the sourceforge discussion
board at
http://sourceforge.net/projects/mpiblast