[GE users] ncbi blast on SGE

reuti reuti at staff.uni-marburg.de
Tue May 25 21:26:53 BST 2010


Zitat von axinsoft <axinsoft at yahoo.com.cn>:

> murple / reuti,
>
> thank you very much for replying. i am very new to SGE, and just   
> realize array jobs may just same copy of jobs running again and again.
>
> i did not complie anything, just use the default "blastall" in Rocks  
>  pack (i am not sure if that is a native one from ncbi).
> in Rocks manaul
> http://www.rocksclusters.org/roll-documentation/bio/5.3/blast_usage.html
> example is very simple and nothing is mentioned.
>
> regarding OpenMPI, i know there is another package "mpiblast" in   
> Rocks too. is that you are refering to? i hear many people   
> recommending mpiblast.
>
> i am just very curious this article is refered by many people(i   
> sawed somewhere like 16 times refer). and did give me very   
> impressive how Grid Engine is working.
> http://developers.sun.com/solaris/articles/integrating_blast.html
>
> it simply use native blast:
> ${progpath}/blastall -p blastn \
>       -d nt     \
>       -i ${seqpath}/nt.123 \
>       -e 0.1    \
>       -o ${outpath}/out.123.blastn
> to get the fancy result of:
>
> job-ID prior name       user   state submit/start at    queue master  
>  ja-task-ID
> ------------------------------------------------------------------------------
> 113    0    blast_5.cs blast   r   05/19/2003 15:57:24 andre.q    MASTER
> 115    0    blast_7.cs blast   r   05/19/2003 15:57:24 andre.q    MASTER
> 116    0    blast_8.cs blast   r   05/19/2003 15:57:24 andre.q    MASTER
> 119    0    blast_11.c blast   r   05/19/2003 15:58:24 andre.q    MASTER
> 120    0    blast_12.c blast   r   05/19/2003 15:58:54 caesar.q   MASTER
> 121    0    blast_13.c blast   r   05/19/2003 15:58:55 saqqara.q  MASTER
> 112    0    blast_4.cs blast   r   05/19/2003 15:57:24 odiche.q   MASTER
> 117    0    blast_9.cs blast   r   05/19/2003 15:57:25 kaiser.q   MASTER
> 118    0    blast_10.c blast   r   05/19/2003 15:57:25 kaiser.q   MASTER
> 122    0    blast_14.c blast   r   05/19/2003 15:59:39 tonylama.q MASTER
> 123    0    blast_15.c blast   qw  05/19/2003 15:57:14
> 124    0    blast_16.c blast   qw  05/19/2003 15:57:14
>
> is that article try to tell me: the program could see SGE by default  
>  a powerful "virtual" computer?

Depends. It can be seen this way, but there is no parallel job running  
in the above example. These are all serial jobs running at their  
native speed. Of course, you will get the total of "all results"  
faster than computing just on one single core machine, as all these  
jobs are running side by side at the same time.

If you want to speed up a single blast job, then mpiblast looks like  
the way to go. I don't know which parallel library was used by  
mpiblst, but when I got you in the right way it seems to be Open MPI.

Then you will need to define a parallel environment (PE) in SGE (check  
the Open MPI website for details), and then it should be possible to  
speed up a single job to deliver the results faster. Be aware, that  
this is often not linear, means doubling the number of used cores will  
not cut the computing time exactly to the half.

-- Reuti


> ------------------------------------------------------
> http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=258467
>
> To unsubscribe from this discussion, e-mail:   
> [users-unsubscribe at gridengine.sunsource.net].
>

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=258520

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list