[Wien] Problems with minimizing forces

L. D. Marks L-marks at northwestern.edu
Thu Jul 22 13:32:30 CEST 2004


Wien is a very nice code, but sometimes it takes a little detective work
to find out what's gone wrong. Some suggestions:

1) Look at the case.output"X" files, where "X" is the last stage when the
calculation crashed, in the cases you showed X=0 or X=1. Sometimes these
have messages in them.

2) Look at the output of the task itself (not :log). For instance, if you
used 'run_lapw -I .... > STDOUT' look in STDOUT. If you ran it as a job,
look in the printout. Sometimes code crashes appear there.

3) Keep a spare copy of Wien with all possible optimizations turned off,
and all possible debug options turned on. For instance, with ifort
-traceback is useful. Change your shell parameter WIENROOT and run it with
this lower level version; sometimes it is more informative. Doing 'man
ifc' or the corresponding for whatever compiler you have will produce a
stack of information (most of it of not much use), and after the reading
it for the 4th or 5th time you'll start to get an idea of what options to
try.

4) Try running it by hand, i.e. the steps (for instance)
x lapw0
x lapw1
....
Sometimes this is clearer. Be aware that you can't always go back. Do a
'save -d My_start' for instance first, a 'clean_lapw' and if you want to
repeat do a 'restore -f -d My_start' and then change something. Out of
frustration I sometimes go back and do the whole init_lapw cycle again
when things are really screwed up; more computer time, less of my time.

5) Remember, finding programming errors is EXPERIMENTAL SCIENCE --
experiment with Wien until you have some partial understanding. Normally
it is operator error, GIGO (Garbage In, Garbage Out).

Good Luck

-----------------------------------------------
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2225 N Campus Drive
Northwestern University
Evanston, IL 60201, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
mailto:L-marks at northwestern.edu
http://www.numis.northwestern.edu
-----------------------------------------------




More information about the Wien mailing list