#BSUB -n 70, but the calculation is slow!!!

option, parallelism,...

Moderators: fgoudreault, mcote

Forum rules
Please have a look at ~abinit/doc/config/build-config.ac in the source package for detailed and up-to-date information about the configuration of Abinit 8 builds.
For a video explanation on how to build Abinit 7.x for Linux, please go to: http://www.youtube.com/watch?v=DppLQ-KQA68.
IMPORTANT: when an answer solves your problem, please check the little green V-like button on its upper-right corner to accept it.
Locked
noro
Posts: 7
Joined: Mon Sep 15, 2014 11:47 am

#BSUB -n 70, but the calculation is slow!!!

Post by noro » Mon Feb 08, 2016 3:44 pm

Good morning,
I do calculations on IBM Platform HPC 4.1.1.1 with OS RHEL 6.4 and PMPI module, with Linux RedHat 6.4 OS.
I have a question about the true number of cores that contribute in the calculation.
Well, I launched a calculating ABINIT and I put in the file script.sh : #BSUB -n 70 , #BSUB -R "span[ptile=10]".
I found that the calculation is slow (More than 10 days for a computation of 4 days), I don’t know is that the calculation is really done on what I requested in script file (n = 70), or he launches on only one or two nodes.

In (lsf2225.o) file, I received the following :
Job was executed on host(s) :
<10*compute028>
<10*compute029>
<10*hpcadmin2>
<10*compute010>
<10*compute011>
<10*compute012>
<10*compute013>

But, at the end of this file (lsf2225.o) :
Resource usage summary:
CPU time : 16.00 sec.
Max Memory : 13 MB (??????????????)
Average Memory : 12.87 MB
Total Requested Memory : -
Delta Memory : -
(Delta: the difference between total requested memory and actual max usage.)
Max Swap : 514 MB
Max Processes : 13 (??????????????)
Max Threads : 14 (??????????????)

Please, let me know if my calculation is really launched on 70 cores and the max RAM, or it is launched on a minimum number of cores.
Thank you very much and best regards.

Locked