
Problem report

This is the description of the ABINIT problem report format.
Later, it will be placed on the Web.
Please refer to the ABINIT Netiquette list as well,
before sending your report via e-mail.

Copyright (C) 2000-2016 ABINIT group (MM,XG)
This file is distributed under the terms of the
GNU General Public License, see ~abinit/COPYING
or http://www.gnu.org/copyleft/gpl.txt .
For the initials of contributors, see ~abinit/doc/developers/contributors.txt .

===============================================

In case of a problem with ABINIT, what can you do ?

First, please refer to ~abinit/KNOWN_PROBLEMS
to see if your problem is reported or not.
If you think that you have found a bug, you
should report it. The best is of course when
you have a bug fix. However usually, you will
ask for help to the ABINIT group.
Note that on the side of the developpers,
there is no obligation to look at your problem,
it is a pure courtesy ... (You know, ABINIT is free ...)

In the case where you have a bug fix,
please send by mail (to gonze@pcpm.ucl.ac.be) :
1) A short description of the problem
  (do not forget the description of the platform,
   if it is platform-dependent, as well as
   the version of code in which this problem appeared)
2) The modified routine(s), in the form of a tar.gz
  file, if there is more than one routine.

If you have a problem without bug fix and ask us to solve it,
we would like you to make a first effort :

- First, examine the log file, to see whether
there might be a WARNING or COMMENT message
suggesting that you
have made a mistake in your input file.

- Second, try to reduce the run so as
to make the problem appear within two minutes on
your machine. Basically, you should reduce ecut,
and reduce the k point sampling. You might as well
try with a smaller cell, if the problem allows it.
If the problem disappear when trying to reduce these,
or if the run becomes meaningless, you should signal it to us.
Note that some bugs seen in production runs at 20 Hartree
or more, can be seen also with 2 Hartree or less !
The physical results might be crazy, but this is not
important if the bug is still visible ...

- Third, take the most debugged version of ABINIT
available for production, and see whether the bug is still
present.

- Fourth, fill the bug report form below, and mail
it to gonze@pcpm.ucl.ac.be .

The bug report format is :

1) Version of ABINIT
   .

2) Machine type : vendor + some hardware characteristics (processor)
   .

3) Operating system (including version number)
   "uname -a" will work on most platforms.
   "oslevel" will be preferable on IBM workstations (AIX).
   .

4) Fortran Compiler & C Compiler(cpp) (including version number)
   For the compiler version number, type man "compiler command" etc.
   e.g.
        pgf90 (PGI compiler on Linux) :  pgf90 -V
        xlf   (on IBM platform)       :  lslpp -l
   You might this information following
   http://www.fortran-2000.com/ArnaudRecipes/Version.html
   Otherwise, ask your system administrator.
   If you use your executables compiled by yourself,
   please give relevant compiler options (or attach your makefile_macros).
   .

5) A short description of the problem
   .
   .
   .

6) Error message, or compiler message ...
   This is mandatory, of course.
   .
   .
   .

7) The time after which the problem occurs.
   If it is larger than two minutes on your machine,
   you must indicate why it was not possible
   to decrease it further thanks to reduction of ecut
   or nkpt, or natom
   .


Moreover, you MUST attach to your mail a tar.gz file
with ALL the input files related to the problem
(input files(s), files file(s), pseudopotential file(s)).
Note that if you have a problem with the response
function part of the code, you must also send the
input file to the prior ground-state calculation.
You might include in the tar.gz file the output or log files.
(But please be careful about your e-mail size.)

