hpc3 클러스터에서 총 87 개 cpu 이용시 예

#PBS -l nodes=10:ppn=8:hpc3+1:ppn=7:hpc3,walltime=100:00:00

또는

#PBS -l nodes=10:ppn=8+1:ppn=7

젠장...이것 하나때문에 얼마나 시간을 많이 보냈는데;;; ㅠㅠ

* 참고자료...

[torqueusers] request odd number of cpus when using ppn=2

• Previous message: [torqueusers] request odd number of cpus when using ppn=2
• Next message: [torqueusers] request odd number of cpus when using ppn=2
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, Oct 21, 2008 at 3:40 PM, Brock Palen <brockp at umich.edu> wrote:



> How could a user request 25 cpus?  For example code has a master process

> that controls several 2 thread MPI ranks.

>

> Thus jobs should be laid out:

>

> 1  2  2  2  2  2   etc,

>

> I tried:

>

> -l nodes=5:ppn=2+nodes=1



-l nodes=5:ppn=2+1:ppn=1

-------------- next part --------------

An HTML attachment was scrubbed...

URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20081021/3fc44fb3/attachment.html

• Previous message: [torqueusers] request odd number of cpus when using ppn=2
• Next message: [torqueusers] request odd number of cpus when using ppn=2
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]