[Bioclusters] Bioinformatics Benchmark System version 3 release candidate 1 ready

Doug Shubert bioclusters@bioinformatics.org
Mon, 12 Jul 2004 14:36:18 -0400

Hello Joe,
Getting a Module error.

[root@opteron bbsv3rc1]# bbsrun --help
Can't locate Object/MultiType.pm in .......

Was this Module compiled with bbs3?


Joe Landman wrote:

>  Much delayed as version 2 went through many rewrites, and we did not
>release early or often.  This will be corrected.
>  http://scalableinformatics.com/BBS/
>  http://bioinformatics.org/bbs
>(they are identical)
>  Much is new in version 3.
>1)  You no longer need to modify program source
>to setup your benchmark.  A simple XML input file will do. 
>2) A number of examples are included.  We need to document the
>format better, though if you read the format, you will see 
>that it really isn't complex (this is what took so long
>oddly enough)
>3) A bbsv1.xml input deck that replicates the original benchmarks from
>last year is included.
>4) New baseline example tests have been created using HMMer, multiple
>BLAST runs, and others.  Further baseline tests will be added (and
>suggestions are always welcome).  We are looking at a number of codes
>including ClustalW, and various chemistry and proteomics codes.
>5) Output in multiple formats: plain text, csv (comma separated values),
>and XML.
>6) A dryrun option.  Does everything but the final execution.  Uses a 
>random sleep rather than a run.  
>Much more code, some of it hinting at things to come.  Quite scalable:
>we ran 100 simultaneous tests on a laptop (learning some surprising
>things about DBM and performance loss in the process).  Documentation
>has been created at least in the form of man pages.  More documentation
>should be available soon, with details of the formats, configuration,
>and so forth
>The code is heavily commented.  It is GPLed for GPL projects/products. 
>For commercial products which need to be non-GPL and supported efforts,
>please contact us.
>Please give it a spin, and send us a note as to how you found it, what
>else you might like on it, and what tests you would like to see.  If you
>want to contribute an idea for a test, or even better, a test, please
>send a note.  If you want to help out on coding, please send a note as