[Pw_forum] Example 06 - problem with phonon calculation

[Axel Kohlmeyer]

On Wed, 18 Feb 2009, Eric Landry wrote:

EL> Paolo,
EL> 
EL> I also have encountered a similar error message when I installed the code on
EL> a machine running Cygwin with the following compilers:
EL> 
EL> MPIF90         = g95
EL> #F90           = g95
EL> CC             = cc
EL> F77            = g95
EL> 
EL> Do you think the compiler might also be the issue here?

the quality of g95 varies a lot with the individual binaries.
the "stable" version seems quite reliable. even though i encountered
a few quirks there as well. the "latest" version of g95 can be
pretty good or pretty bad. 

mind you, the code that i am using g95 on, is using "modern" fortran 
90/95 extensively, so it is exercising compilers hard, much, much more
than Q-E does. gfortran currently doesn't stand a chance, PGI is 
a continuous catastrophe requiring many workarounds, and intel is 
a hit and miss (different compiler versions and patchlevels 
miscompile different parts of the code, however, most of the time
only with full optimization). for this code, the stable g95 has
the least problems. so i'd assume the same would hold for Q-E,
unless there is some specific code construct used somewhere in Q-E, 
that g95 doesn't like (happens every once in a while).

hope that helps (even though it is not very definite...),
   axel.


EL> 
EL> Thanks,
EL> Eric
EL> 
EL> -----Original Message-----
EL> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
EL> Behalf Of Paolo Giannozzi
EL> Sent: Wednesday, February 18, 2009 9:18 AM
EL> To: PWSCF Forum
EL> Subject: Re: [Pw_forum] Example 06 - problem with phonon calculation
EL> 
EL> Paolo Giannozzi wrote:
EL> > Eric Landry wrote:
EL> > 
EL> >> My make.sys file says this:
EL> >>
EL> >> MPIF90         = mpif90
EL> >> #F90           = gfortran
EL> >> CC             = gcc
EL> >> F77            = gfortran
EL> > 
EL> > most likely you have gfortran as underlying compiler in mpif90.
EL> > Only recent versions of gfortran
EL> 
EL> ...produce executables that work, and maybe only part of them actually works
EL> (sorry for the previous truncated message)
EL> 
EL> Paolo
EL> --
EL> Paolo Giannozzi, Democritos and University of Udine, Italy
EL> _______________________________________________
EL> Pw_forum mailing list
EL> Pw_forum at pwscf.org
EL> http://www.democritos.it/mailman/listinfo/pw_forum
EL> 
EL> 
EL> _______________________________________________
EL> Pw_forum mailing list
EL> Pw_forum at pwscf.org
EL> http://www.democritos.it/mailman/listinfo/pw_forum
EL> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.