crash in mod_timing.f90

Various technical topics such as parallelism and efficiency, netCDF problems, the Yambo code structure itself, are posted here.

Moderators: Daniele Varsano, andrea.ferretti, andrea marini, Conor Hogan, myrta gruning

crash in mod_timing.f90

Postby martin.gmitra » Mon Sep 23, 2013 9:21 am

Dear yambo developers,

I am experiencing yambo crash with the following error message:
Code: Select all
At line 229 of file mod_timing.f90
Fortran runtime error: Unexpected element '*' in format
(a,**x,a)
   ^


I think this might be related with the nodes numbering which in that case
goes in parallel over 100. From the last hint in the "l-*" output file I can deduce that some
nodes "**" finished calculations:
<12h-06m-36s> P**: G0W0 PPA |####################| [100%] 12h-05m-46s(E) 12h-05m-46s(X)

Thank you in advance for any hints.
Best regards,
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Re: crash in mod_timing.f90

Postby Daniele Varsano » Mon Sep 23, 2013 9:57 am

Dear Martin,
what version of Yambo are you using?
Anyway the problem looks related as you say to the format. You can try to substitute the i2.2 in the mode_timing module
with i3.3. May someone else can give you a better workaround.

Cheers,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
User avatar
Daniele Varsano
 
Posts: 2060
Joined: Tue Mar 17, 2009 2:23 pm

Re: crash in mod_timing.f90

Postby martin.gmitra » Mon Sep 23, 2013 10:18 am

Dear Daniele,

Thanks for your replay. I am running devel version, revision 48.
In the mod_timing.f90 there are several hash hints, and I am not sure
If I am able to change all of them consistently.

Best,
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Re: crash in mod_timing.f90

Postby Daniele Varsano » Mon Sep 23, 2013 11:34 am

Dear Martin,
the source has been modified, update your source using
Code: Select all
svn up

recompile and let us know if your problem is solved.

Best,

Daniele
Dr. Daniele Varsano
S3-CNR Institute of Nanoscience and MaX Center, Italy
MaX - Materials design at the Exascale
http://www.nano.cnr.it
http://www.max-centre.eu/
User avatar
Daniele Varsano
 
Posts: 2060
Joined: Tue Mar 17, 2009 2:23 pm

Re: crash in mod_timing.f90

Postby martin.gmitra » Mon Sep 23, 2013 1:35 pm

Dear Daniele,
Thanks for the update. It seems that job successfully restarted. I have a small request,
could you also implement the change to src/io/Fragments_Restart.F file as we discussed
in http://www.yambo-code.org/forum/viewtopic.php?f=10&t=701 post?

Thanks for your help,
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Re: crash in mod_timing.f90

Postby martin.gmitra » Mon Sep 23, 2013 1:51 pm

I am sorry! I have been too fast. The error
Code: Select all
At line 229 of file mod_timing.f90
Fortran runtime error: Unexpected element '*' in format
(a,**x,a)
   ^

is still there!
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Re: crash in mod_timing.f90

Postby andrea marini » Mon Sep 23, 2013 2:10 pm

This is strange. Did you do a

Code: Select all
>make clean_all


first ? With how many cpu are you trying to run ? Can you run with few cpu's ?

Andrea
Andrea MARINI
Istituto di Struttura della Materia, CNR, (Italy)
User avatar
andrea marini
 
Posts: 447
Joined: Mon Mar 16, 2009 4:27 pm

Re: crash in mod_timing.f90

Postby martin.gmitra » Mon Sep 23, 2013 8:16 pm

Yes, I did clean_all, recompiled yambo from scratch and run using 80 cores.
Job restarted, completed several thousands of sections and crashed after
five hours.
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Re: crash in mod_timing.f90

Postby andrea marini » Mon Sep 23, 2013 8:57 pm

Is it the crash still related to mod_timing ? If yes can you provide all relevant input files (GS and Yambo) ? But we need first to reproduce the error with a smaller run. Can you try to reduce the parameters as much as possible in order to reproduce the error ?

Andrea
Andrea MARINI
Istituto di Struttura della Materia, CNR, (Italy)
User avatar
andrea marini
 
Posts: 447
Joined: Mon Mar 16, 2009 4:27 pm

Re: crash in mod_timing.f90

Postby martin.gmitra » Fri Oct 18, 2013 3:54 pm

Dear all,
Finally I have realized that troubles were due to interrupted y2p run (walltime limit on cluster).
Continuation somehow provides ill databases. Redoing databases from scratch I have been
able successfully finish the calculations.

Thanks for your participation,
Martin Gmitra, PhD
University of Regensburg, Germany
martin.gmitra
 
Posts: 26
Joined: Tue Apr 23, 2013 8:40 pm

Next

Return to Technical Issues

Who is online

Users browsing this forum: No registered users and 1 guest