[Bioclusters] questions with mpiBLAST

[Jeremy Mann]

I asked these questions about 2 weeks ago. I need answers before I
implement this system wide.

One, since mpiformatdb splits up each database according to the -N
value, why does each node have several of these in its local storage?

Two, why the constant transferring of the database segments to
individual nodes? If there is an explanation or a reason for this, I'd
love to hear it.

Three, one and two mess things up when you have multiple users trying to
use the system. Obviously the segments get copied to local storage owned
by THAT user. If another user runs mpiblast, I get permission denied
errors because it tries to copy over the existing users database
segments.

I can get around this by using --removedb, but the question still
remains, WHY formatdb into segments for every node IF you still have to
remove the database so another user can use it?

It seems to me, we are still stuck with the age old NFS problem. Even
though the dbs are stored locally, each time you run mpiblast it still
copies segments to available nodes. What is different between this and
using normal shared storage? Your still using the network in both cases,
albeit for mpiblast your doing a much smaller 'segment' of that
database.


Thanks for any answers I can get ;)



-- 
Jeremy Mann
jeremy@biochem.uthscsa.edu

University of Texas Health Science Center
Bioinformatics Core Facility
http://www.bioinformatics.uthscsa.edu
Phone: (210) 567-2672


-- 
Jeremy Mann
jeremy@biochem.uthscsa.edu

University of Texas Health Science Center
Bioinformatics Core Facility
http://www.bioinformatics.uthscsa.edu
Phone: (210) 567-2672