Hello! I have at least a couple blastn queries that are bogging down in readPartialBioseqs(), and I was wondering how I could get around this problem. It looks like these queries are generating many partial results per database fragment (e.g., 20K partial results per fragment). As a result, a blastall job that takes less than a minute ends up taking about 40 minutes with mpiblast. Very strange!
Attached is one of those queries, which is actually from this NCBI benchmark suite: ftp://ftp.ncbi.nlm.nih.gov/blast/demo/benchmark/benchmark.tar.gz An 8-worker blastn job with default options bogs down with either the nt database or the smaller benchmark.nt database in the benchmark suite above. Setting a smaller -b and -v option improves performance, but I would like to be able to run blastn with the default options. I also tried --disable-mpi-db, but this seemed to have no effect. Any advice? I see this behavior with the latest mpiblast tarball and ncbi blast 2.2.14 on both our Altix (IA64) and Altix XE (x86_64) systems, both running SuSE Linux (SLES9) and compiled using the Intel 9.1 compilers and Intel MPI libraries. Thanks so much! -- Haruna @ SGI
NM_018073
Description: Binary data
------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________ Mpiblast-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mpiblast-users
