In article <4cu6lr$a73 at larry.rice.edu>, fant at babs.rice.edu says...
>>Has anyone ever configured GCG to use NQS as the batch mechanism for
long
>sequence searches and alignments under GCG? Unix batch is pretty much
useless
>for any sort of serious load balancing and scheduling, and I am in the
>process of setting up NQS or some decendent thereof to handle scheduling
>on our local cluster. I would like to use it for long GCG runs as well,
>if anyone can make any suggestions as to how to do this.
>>Thanks in advance,
> Andy
>Andy,
I run both GCG and NQS here (on a Dec2100 under DU), and there is no need
to "configure" GCG to run in the batch mode with NQS (I agree with you
that the Unix batch is not a batch...)
Suppose I want to run FASTA. What I do is:
- create a file named "fasta.sub" which will be some sort of:
#-----------------------------------
source /my_gcgdir/gcgstartup
gcg
/my_gcgdir/gcgbin/execute/fasta -nomon -opt -align=10 << END
my_query_seq
embl:*
6
100
my_output_file
0
END
#---------------------------------------
And then you just type "qsub fasta.sub" and here it goes....
In fact, for more interactivity and more friendliness to the others, the
fasta.sub file is created by a simple Perl script that asks the user what
he (she) wants.
Go for NQS. It's just great, and a REAL batch!!
Good luck,
Jean-Loup
--
-------------------------------
Jean-Loup RISLER
risler at genetique.uvsq.fr
Universite de Versailles
45 Avenue des Etats-Unis
78035 Versailles France
-------------------------------
