[Wien] Patch to fix ROTDEF error in lapw0 for Fd3m et al.

[Peter Blaha]

Thank you very much for reporting this issue and the fix.


The funny thing is, that rotdef in SRC_dstart runs without problem, 
although the part of the code for finding the symmetry operations is 
exactly the same.

What is different, is that in dstart the rotdef.f routine does NOT have 
a symmerization of the positions, while the scf-programs (lapw0/1,2,so) 
have an additional symmetrization check of equivalent positions. This 
check, however, leads to a small modification due to some rounding error 
in the 16th digit in the positions, and for those atoms, the following 
code fails, because the mod(x,one) gives still "one" instead of "zero".


Thus there are 2 possible fixes:  your patch, and/or removing this 
symmetrization. I intend to apply both fixes, because this 
symmetrization should not be in the scf programs anyway.


Thank you very much for the report.

Am 23.01.2021 um 20:02 schrieb Martin Kroeker:
> We have recently encountered an unexpected problem with lapw0 returning
> "ROTDEF error", claiming "symmetry operation not found" for equivalent
> positions of an X,X,X site in spacegroup Fd-3m. (One simplified example
> is attached to this message to allow reproducing the problem.)
> As it appears, SRC_lapw0/rotdef.f takes turns applying the centering
> operations to the calculated distance vector looking for a match - and
> then undoes that operation before trying the next, instead of simply
> copying back the original value. This demonstrably leads to rounding
> errors that can exceed the hardcoded tolerance threshold of 1.5e-7
>
> The attached patch avoids the problem by trivially saving x1 before
> the cascade of tests for centered lattices, and restoring it after
> each that did not lead to a successful exit from the loop.
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:  http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html

-- 
-----------------------------------------------------------------------
Peter Blaha,  Inst. f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-158801165300
Email: peter.blaha at tuwien.ac.at
WWW:   http://www.imc.tuwien.ac.at      WIEN2k: http://www.wien2k.at
-------------------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20210126/0bfe57fb/attachment.htm>