[Wien] LAPW1 error

Laurence Marks laurence.marks at gmail.com
Mon Oct 5 21:36:36 CEST 2020


Which version of ifort (do ifort -version)? I know there is a bug with some
2017 and -assume buffered_io that effects the mixer. It might be connected,
or not.

Ifort can be buggy, but it is faster than gfortran with Intel CPUs. It is
good to keep more than one version so one can regress to check problems.
Also the Intel IMPI is buggy with a memory leak that was only recently
fixed.

_____
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

On Mon, Oct 5, 2020, 07:23 Israel Omar Perez Lopez <israel.perez at uacj.mx>
wrote:

> Hi
>
> That same problem happens to me with version 19.1 and 19.2 for some basic
> structures (for instance Li Im-3m) using ifort 2018, that doesn't happen
> with gfortran for version 18 or 19. I do not know why. But when it happens,
> I use the .machines file and set the omp_global to 1 and run in parallel.
> The error no longer appears. Apparently, there is a problem with the
> OMP_NUM_THREADS variable. If during installation of Wien2k the system
> detects 2 cores, but you set it 4 or 8, those kinds of errors show up. In
> my case, I have 4 cores in a laptop. If I set it to 4 the problem shows up
> every time a SCF is executed. Then I set to 2 and works fine for most
> structures but in some fails. In those cases I create the .machines and set
> omp_global to 1 and no error is issued.
>
> Hope this helps
>
> Dr. Israel Pérez
>
> Institute of Engineering and Technology
>
> Department of Physics and Mathematics,
>
> Universidad Autónoma de Ciudad Juárez
>
> Av. del Charro 450 Nte., Col. Partido Romero,
>
> Ciudad Juárez, Juarez Chihuahua. Mexico C. P. 32310
>
> Tel: +52 (656) 688 4887
>
>
> National Council of Science and Technology
>
> Insurgentes Sur No. 1582,
>
> Col. Crédito Constructor,  C.P. 03940
>
> Del. Benito Juárez, México D. F.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20201005/81ef7528/attachment.htm>


More information about the Wien mailing list