Re: mpiblast

Hi Chuming,

How much of that 17 seconds is startup time? -- if you run with a single short query on the database how long
does mpiblast take?
You're probably reaching a lowerlimit for time at very small query sizes. How many nodes are you running on?
How long does it take to run an mpi-hello_world on that number of nodes? 
If you'd like you can post or look through the output produced when running with --debug and see where the
most time is spent for the short run. 

additionally, please don't double-post to mpiblast-users and bioclusters. A question like this is for
the mpiblast list.
-Lucas

On Wednesday, March 15, 2006 at 08:17 -0500, Chuming Chen wrote:
> Hi all,
> 
> Does the query size of mpiblast effect the total performance?
> 
> It took about 17 seconds per sequence when I run 721  sequences  (542K) 
> against 22G database.
> But when I run 88532 sequences (47M) against the same database, it  took 
> about 1 minute.
> 
> Can the performance be improved if the query sequence is in a  
> relatively smaller size?
> 
> Thank you for your kind comments and suggestions.
> 
> Chuming Chen