[Wien] configuring static linking with siteconfig
Laurence Marks
L-marks at northwestern.edu
Mon Oct 30 16:21:55 CET 2017
I am going to beg to differ about dynamic versus static compiling. It is
convenient to have a version that can be ported to other clusters and only
need dynamic linking of the mpi fabric (i.e. impi or openmpi). Not every
cluster has up to date (or even working) scalapack!
The same holds for fftw...
It should not be hard to do something like add a "ifndef $SCALAPACK" to the
Makefiles, for version 17.2 (or 18.1).
(The user is right, except when she is wrong.)
On Mon, Oct 30, 2017 at 10:03 AM, Ruh Thomas <thomas.ruh at tuwien.ac.at>
wrote:
> Dear Pavel,
>
> I looked into static linking with no / minimal changes into the default
> Makefiles of WIEN2k and I am afraid it is not possible - the suggested
> workaround by Gavin should not work. At least it did not for me. Mainly
> because, as Gavin already mentioned, siteconfig_lapw sets all the linking
> up dynamically.
> Our point of view in the WIEN-group is that dynamic linking is
> preferential anyway (especially if you want mpi-parallelism), because the
> binaries would be rather large.
>
> In case you absolutely need statically linked ScaLAPACK and/or
> FFTW-libraries you will have to manually adapt the necessary Makefiles.
>
> Kind regards
> Thomas
>
> ________________________________________
> Von: Wien <wien-bounces at zeus.theochem.tuwien.ac.at> im Auftrag von Gavin
> Abo <gsabo at crimson.ua.edu>
> Gesendet: Montag, 30. Oktober 2017 12:59
> An: wien at zeus.theochem.tuwien.ac.at
> Betreff: Re: [Wien] configuring static linking with siteconfig
>
> In WIEN2k versions older than 17.1, I think the advanced user could
> directly modify WIEN2k_OPTIONS (or OPTIONS), then run siteconfig to load
> the file and just do a save to propagate the settings to the Makefiles.
> I think it might be with the parallel settings that the new
> auto-generation in siteconfig_lapw changed it so that also no longer works.
>
> On line 2036 in siteconfig_lapw, there should be:
>
> set SCALAPACK_LIBNAME = 'mkl_scalapack_lp64'
>
> Have you tried changing it to:
>
> set SCALAPACK_LIBNAME = 'ibscalapack.a'
>
> Similarly, maybe try changing line 2795:
>
> set fftwlibspara = "-l${FFTW_LIBNAME}_mpi"
>
> to say
>
> set fftwlibspara = "-lfftw3_mpi.a"
>
> On 10/30/2017 3:21 AM, Pavel Ondračka wrote:
> > Dear Wien2k mailing list,
> >
> > what is the recomended way to link Wien2k with static libs using the
> > siteconfig script? For example when I go to "SCALAPACK Settings" I
> > would like to link with static libscalapack.a, however I can only set
> > the SCALAPACKROOT and SCALAPACK_LIBNAME and the final SCALAPACK_LIBS is
> > autogenerated to link with a dynamical library. Is there a way to edit
> > the SCALAPACK_LIBS and link with a static library without editing the
> > generated makefiles manually? BTW similar problems is with fftw
> > settings...
> >
> > Best regards
> > Pavel Ondračka
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.
> theochem.tuwien.ac.at_mailman_listinfo_wien&d=DwIFBA&c=
> yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_
> T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Q6Vgp8k1XOuU6scVd5mmvZz1YmYubE
> 5qd9eDgXDOH9g&s=xSQeLsy-LVR66lvZdADYnonOoM0A44mTqV0kKpZScZA&e=
> SEARCH the MAILING-LIST at: https://urldefense.proofpoint.
> com/v2/url?u=http-3A__www.mail-2Darchive.com_wien-
> 40zeus.theochem.tuwien.ac.at_index.html&d=DwIFBA&c=
> yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_
> T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Q6Vgp8k1XOuU6scVd5mmvZz1YmYubE
> 5qd9eDgXDOH9g&s=wdlvciFGzC1kPAxeB-rTSyqpzfXnq5o5X9P5ynYqpW4&e=
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.
> theochem.tuwien.ac.at_mailman_listinfo_wien&d=DwIFBA&c=
> yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_
> T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Q6Vgp8k1XOuU6scVd5mmvZz1YmYubE
> 5qd9eDgXDOH9g&s=xSQeLsy-LVR66lvZdADYnonOoM0A44mTqV0kKpZScZA&e=
> SEARCH the MAILING-LIST at: https://urldefense.proofpoint.
> com/v2/url?u=http-3A__www.mail-2Darchive.com_wien-
> 40zeus.theochem.tuwien.ac.at_index.html&d=DwIFBA&c=
> yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_
> T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=Q6Vgp8k1XOuU6scVd5mmvZz1YmYubE
> 5qd9eDgXDOH9g&s=wdlvciFGzC1kPAxeB-rTSyqpzfXnq5o5X9P5ynYqpW4&e=
>
--
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what nobody
else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu ; Corrosion in 4D: MURI4D.numis.northwestern.edu
Partner of the CFW 100% program for gender equity, www.cfw.org/100-percent
Co-Editor, Acta Cryst A
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20171030/37a34268/attachment.html>
More information about the Wien
mailing list