[gpaw-users] NaN atomic radii in solvation model

[Alexander Held]

Dear GPAW users/developers,

 

I recently found that the continuum solvation model silently ignores NaN
values for the atomic radii. This could affect calculations involving
elements where ASE's van der Waals radius is NaN. Such calculations possibly
produce finite but nonsensical output. Please check your calculations if you
have been using the solvation model.

 

I have opened issue https://gitlab.com/gpaw/gpaw/issues/53 and there is
already a fix available as merge request
https://gitlab.com/gpaw/gpaw/merge_requests/173 . An exception is now raised
when one tries to use NaN atomic radii and also the root cause of the
problem (lazy floating point error masking) has been fixed in the merge
request.

 

I created the feature branch some days ago and now gpaw/gpaw/master is ahead
of it. I am not sure if I should merge gpaw/gpaw/master back into my feature
branch in this case or not (I did not do it).

 

Best regards

Alexander Held

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/gpaw-users/attachments/20170209/c33bc1d9/attachment.html>