[ase-users] EOF Error Vibrational Modes Espresso

Thu Dec 7 01:29:44 CET 2017

As a brief aside:

With regards to Ask's comment about the Vibrations class, I have also found
that ASE cannot parallelize the vibrational mode calculations when using
mpi4py and the Vasp calculator (haven't tried the Espresso calculator yet).
If I recall correctly, it never gets passed the first vibrational mode
successfully.

Andrew

On Wed, Dec 6, 2017 at 6:24 PM Lance Kavalsky via ase-users <
ase-users at listserv.fysik.dtu.dk> wrote:

> Hi Ask,
>
>
> Your suggestion fixed it!
>
>
> Thanks so much for you help.
>
>
> Best,
>
> Lance
> ------------------------------
> *From:* Ask Hjorth Larsen <asklarsen at gmail.com>
> *Sent:* Saturday, December 2, 2017 7:06:32 PM
>
> *To:* Lance Kavalsky
> *Cc:* ase-users at listserv.fysik.dtu.dk
> *Subject:* Re: [ase-users] EOF Error Vibrational Modes Espresso
> Hi.
>
> 2017-12-02 23:12 GMT+01:00 Lance Kavalsky <lance.kavalsky at mail.utoronto.ca
> >:
> > Hello,
> >
> >
> > Below is the traceback. It reaches this point and then hangs. I am
> running
> > the script with mpirun on 4 processors (mpirun -np 4 python
> trans_ase.py).
> >
> >
> > As for the pseudopotentials, I am using Na.pbe-mt_fhi.UPF and
> > P.pbe-mt_fhi.UPF which were obtained from the Quantum Espresso webpage.
> >
> >
> > Writing vib.log.0x+.pckl
> > Writing vib.log.eq.pckl
> > Writing vib.log.0y-.pckl
> > Writing vib.log.0x-.pckl
> > Writing vib.log.0y+.pckl
> > Writing vib.log.0z+.pckl
> > Writing vib.log.1x-.pckl
> > Writing vib.log.0z-.pckl
> > Writing vib.log.1z-.pckl
> > Writing vib.log.1y+.pckl
> > Writing vib.log.1y-.pckl
> > Writing vib.log.1x+.pckl
> > Traceback (most recent call last):
> >   File "trans_ase.py", line 23, in <module>
> >     vib.summary(log='vib.txt')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 292, in summary
> >     hnu = self.get_energies(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 256, in get_energies
> >     self.read(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 210, in read
> >     fplus = load(name + '+.pckl')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 195, in load
> >     f = pickleload(fl)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 1384, in load
> > Traceback (most recent call last):
> >   File "trans_ase.py", line 23, in <module>
> >     vib.summary(log='vib.txt')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 292, in summary
> >     hnu = self.get_energies(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 256, in get_energies
> >     self.read(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 210, in read
> >     fplus = load(name + '+.pckl')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 195, in load
> >     f = pickleload(fl)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 1384, in load
> > Traceback (most recent call last):
> >   File "trans_ase.py", line 23, in <module>
> >     vib.summary(log='vib.txt')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 292, in summary
> >     hnu = self.get_energies(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 256, in get_energies
> >     self.read(method, direction)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 210, in read
> >     fplus = load(name + '+.pckl')
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/site-packages/ase/vibrations/vibrations.py",
> > line 195, in load
> >     f = pickleload(fl)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 1384, in load
> >     return Unpickler(file).load()
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 864, in load
> >     return Unpickler(file).load()
> >     return Unpickler(file).load()
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 864, in load
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 864, in load
> >     dispatch[key](self)
> >     dispatch[key](self)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 886, in load_eof
> >     dispatch[key](self)
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 886, in load_eof
> >   File
> >
> "/bg01/homescinet/c/csingh/kavalsky/CONDA_ENVS/ase/lib/python2.7/pickle.py",
> > line 886, in load_eof
> >     raise EOFError
> >     raise EOFError
> > EOFError
> >     raise EOFError
> > EOFError
> > EOFError
>
> This contains repetitive lines.  I infer that it ran on multiple
> cores, and that probably the simultaneous opening of the same files
> lead to garbled data.
>
> I believe the Vibrations class should work in parallel, so this is
> probably a bug.  However it is somewhat unusual to run ASE scripts in
> parallel except with GPAW and ASAP, so I suspect you should be doing
> something else.
>
> Did you run this in e.g. slurm through "srun python ...."?  Perhaps
> what you want is to not put the srun command in the submit script, but
> rather inside the Python script:
>
>   calc = Espresso(command='srun my-espresso-command', ...)
>
> or equivalently export ASE_ESPRESSO_COMMAND="srun my-espresso-command".
>
> That way ASE will run in serial, but invoke QE in parallel.
>
> ASE may run indeed correctly in parallel with mpi4py, but it appears
> to me that the Vibrations class is not quite fit for the task at the
> moment.
>
> Best regards
> Ask
>
> >
> >
> >
> > Thanks,
> >
> > Lance
> >
> >
> >
> > ________________________________
> > From: Ask Hjorth Larsen <asklarsen at gmail.com>
> > Sent: Saturday, December 2, 2017 7:42:51 AM
> > To: Lance Kavalsky
> > Cc: ase-users at listserv.fysik.dtu.dk
> > Subject: Re: [ase-users] EOF Error Vibrational Modes Espresso
> >
> > Hi,
> >
> > 2017-12-01 18:39 GMT+01:00 Lance Kavalsky via ase-users
> > <ase-users at listserv.fysik.dtu.dk>:
> >> Hello,
> >>
> >>
> >> My end goal is to calculate the vibrational energies (and subsequently
> ZPE
> >> and TS) of an adatom on a fixed substrate using ase.vibrations.
> >>
> >>
> >> As a test case, I have been trying to reduce my system to only 2 atoms
> (1
> >> P
> >> and 1 Na) to ensure my script is functioning properly (attached). The
> >> script
> >> successfully reads from the Quantum Espresso input file, and vib.run()
> >> will
> >> write the pckl displacement files, but it returns EOFError when trying
> to
> >> execute vib.summary().
> >>
> >>
> >> Any advice or guidance would be very much appreciated.
> >
> > I don't know if I can help with vibrations, but in case someone can,
> > could you give the full stack trace please?  It would be useful to
> > know what it was reading when it caused EOF.
> >
> > The pseudopotentials and version numbers might also prove useful.
> >
> > Best regards
> > Ask
> >
> >>
> >>
> >> Thanks,
> >>
> >> Lance Kavalsky
> >>
> >> HBSc., MASc. Student
> >>
> >> Dept. of Materials Science and Engineering
> >>
> >> University of Toronto
> >>
> >>
> >>
> >> _______________________________________________
> >> ase-users mailing list
> >> ase-users at listserv.fysik.dtu.dk
> >> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20171207/2a3c82be/attachment-0001.html>