Mike Mettke wrote:
Jose,

pbs thinks each of your machines has only one cpu.
Please execute for each node:

qmgr -c "s n neurona1.brain np=2"
Well, actually I did:
qmgr -c "s n neurona1.brain np=8"
because there 8 client nodes and I'd like to have at least 8 parallel processes running in the cluster.

All right! It's now working properly.

Anyway, this solution doesn't scale up easily. I mean, I would have to change the definitions for every node everytime a new node is added/deleted. Maui/PBS still allocate all of the nodes to every new job submitted:
----
/opt/maui/bin/checkjob 40


checking job 40

State: Running (User: jmbs Group: difuso Account: [NONE])
WallTime: 1:41:03 (Limit: INFINITY)

QueueTime: Wed Dec 11 18:38:09
StartTime: Wed Dec 11 18:38:09

Total Tasks: 8

Req[0] TaskCount: 8 Partition: DEFAULT
Network: [NONE] Memory >= 0 Disk >= 0 Features: [NONE]
Opsys: [NONE] Arch: [NONE] Class: [workq 1]

Allocated Nodes:
[neurona8.brain:1][neurona7.brain:1][neurona6.brain:1][neurona5.brain:1]
[neurona4.brain:1][neurona3.brain:1][neurona2.brain:1][neurona1.brain:1]


IWD: [NONE] Executable: [NONE]
QOS: DEFAULT Bypass: 0 StartCount: 1
Partition Mask: [ALL]
Flags: RESTARTABLE

Reservation '40' (-1:40:46 -> INFINITY Duration: INFINITY)
PE: 8.00 StartPriority: 1201

----

Is there anyway to make them allocate only one processor for each new process?
I've read something related to this on the qsub man page (option -l resource_list). But I don't know to config the system to get it done. Just for your reference, the jobs I'm submitting are scripts like this (b.sh):
----
#!/bin/tcsh
#PBS -N BA3
#PBS -o salida3txt
#PBS -e errores3.txt
#PBS -q workq

set program = ~/jl/ba/ba
set prefix = ~/jl
set seed1 = 12264671;
set iter = 6000;


$program $prefix/datos_sc/iono.tra $prefix/datos_sc/iono.tst $iter $seed1
----

And I submit them with a single: "qsub b.sh"



--
Jose Manuel Benitez Sanchez e-mail: [EMAIL PROTECTED]
http://decsai.ugr.es/~jmbs
Dpto. de Ciencias de la Computacion e Inteligencia Artificial
E.T.S. Ingenieria Informatica
Universidad de Granada Tel. +34 - 958 - 24 61 43
18071 - GRANADA (Spain) Fax: +34 - 958 - 24 33 17




-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
_______________________________________________
Oscar-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/oscar-users

Reply via email to