[Wien] lapw1 and lapw2 not running in parallel

[pboulet]

Dear all,

I have compiled Wien2K with openmpi 1.6.3, intel fortran 13.0 and FFT 3.3. The scheduler is OAR.

It seems to work fine for the example files CoS2 except that lapw1 and lapw2 « refuse » to run in parallel. Here is the dayfile content for the last cycle:
    cycle 11 	(Wed Feb 21 15:26:49 CET 2018) 	(30/89 to go)
>   lapw0  -p	(15:26:49) starting parallel lapw0 at Wed Feb 21 15:26:49 CET 2018
-------- .machine0 : 12 processors
30.073u 1.343s 0:04.38 717.1%	0+0k 0+16io 0pf+0w
>   lapw1  -p    	(15:26:53) starting parallel lapw1 at Wed Feb 21 15:26:54 CET 2018
->  starting parallel LAPW1 jobs at Wed Feb 21 15:26:54 CET 2018
running LAPW1 in parallel mode (using .machines)
running lapw1 in single mode
61.065u 0.857s 1:01.94 99.9%	0+0k 0+16io 0pf+0w
>   lapw2 -p     	(15:27:55) running in single mode
19.541u 0.525s 0:20.09 99.8%	0+0k 0+8io 0pf+0w
>   lcore	(15:28:16) 0.012u 0.003s 0:00.02 50.0%	0+0k 0+8io 0pf+0w
>   mixer	(15:28:16) 0.048u 0.007s 0:00.06 66.6%	0+0k 0+8io 0pf+0w
:ENERGY convergence:  1 0.0001 .0000634500000000
:CHARGE convergence:  1 0.01 -.008733
ec cc and fc_conv 1 1 1

>   stop

Obviously lapw1 and lapw2 are running in single mode.

In the manual it is said that during the lapw1para run the .processes file is created. In my case, it is not.

For information, here is my .machines file content:
lapw0: node043:12
dstart: node043:12
nlvdw: node043:12
granularity:1
extrafine:1
lapw2_vector_split:4

Could someone give me a hint on what is going on?

Thank you
Best regards


Pascal Boulet
—
Professor in computational chemistry - DEPARTMENT OF CHEMISTRY
Director of the Madirel laboratory
Aix-Marseille University - Avenue Escadrille Normandie Niemen - F-13013 Marseille - FRANCE
Tél: +33(0)4 13 55 18 10 et +33(0)4 13 55 18 19 - Fax : +33(0)4 13 55 18 50
Email : pascal.boulet at univ-amu.fr <mailto:pascal.boulet at univ-amu.fr>








-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20180221/2b48b1f8/attachment.html>