Page 1 of 1

Bug in Martins2Abinit?

Posted: Mon Jun 21, 2010 6:40 pm
by jschrier
I have downloaded the Martins2Abinit pseudopotential generation code listed on the Abinit website http://www.abinit.org/downloads/psp-links/psp-links/MARTINS2ABINIT.tgz

Compiled/ran the atom.exe code without errors. Likewise compiled the Martins2Abinit.f90 code. However, when I run it, it
does not actually output the "zatom, zion, pspdat" line. Also the format of the "pspcod, pspxc, lmax" line is all screwed up.

Has anyone encountered this before? Is this a known error?

Re: Bug in Martins2Abinit?

Posted: Fri Jun 25, 2010 1:35 pm
by torrent
From my knowledge, such an issue was not encountered before.

The corresponding line (nr 460 in Martins2Abinit.f90):

Code: Select all

write(7,'(f3.0,2x,f5.2,2x,a6,30x,a)') zz,zval,convertdate(strdat(1:9))," : zatom,zion,pspdat"

is present.
There is perhaps a difficulty, on your architecture, with the character manipulation.
You could try to replace this line by:

Code: Select all

write(7,'(f3.0,2x,f5.2,2x,a6,30x,a)') zz,zval,"000000"," : zatom,zion,pspdat"

(put "000000" for the date...)