The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

On 11 Nov 2006, at 09:16, Anglade Pierre-Matthieu wrote:
> If you want more safety, you must not rely fully on the automatic test
> analysis. You'd better look at the output of each failed tests and
> check what the failing reasons was. There is a lot of reasons that may
> produced a FAILED in the log while you'd consider
> that the test successfull or at least not a problem.

Dear Manolo,

Indeed, as mentioned by Pierre-Matthieu, this situation (ABINIT declares
that the test failed, while for usual production there is no  
problem)  might still
be quite frequent. This analysis (succeeded/passed/failed) is a new  
feature
of v5.2 and needs some fine tuning. If you worry about the failed
test cases, you should do three things :
- look in the README files (e.g. ~abinit/tests/v1/README) the  
description
 of the test case that failed (may be you do not care about this  
functionality)
- look in the other files that have been produced by the automated  
testing procedure
 (the output file, the diff file with respect
 to the reference, the fldiff.report file, and finally the summary  
file - some of these
 are described in the installation_notes), maybe the problem seen by  
the
 analysis is at a level not relevant for your work
- look into the ~abinit/KNOWN_PROBLEMS file, as there might be  
comments about problems
 existing on many platforms (e.g. for v5.2.3, in P52.1, it is  
mentioned that Test_v2#15, Test_v2#48 and
 Tutorial#elphon_5  are fluctuating too much for a good portability,  
but there is no real
 problem for production)

Xavier