[Pw_forum] small bug zmesh=0

Fernando A Reboredo reboredofa at ornl.gov
Tue Sep 19 23:23:52 CEST 2006


I found a problem in PP/pw2casino zmesh is set to zero



Syntoms

Line 262 of PP/pw2casino.f90



WRITE(io,'(i6,3f20.14)') INT(zmesh(ityp(na))), (alat*tau(j,na),j=1,3)



Is supposed to write the atomic number  Z of the atom "na".



However the vector zmesh (defined in Modules/atom.f90) is zero.



It seams that my problem is that my pseudopotential is in UPF format. It 
appears to me that in that format the only way to set values for zmesh is to 
include a

<PP_ADDNFO> tag with additional parameters in the pseudopotential file

According to the soubroutine read_pseudo_upf,  zmesh appears to be read 
under that tag. But if i add the <PP_ADDINFO> tag I will leave

upf%has_so=.true.

And I would also have to set spin-orbit information into the pseudopotential 
file with uncertain outcome for me.



The quit fix for me has been to change pw2casino in the following way



Add atm to line 63 that is

USE ions_base, ONLY : nat, ntyp => nsp, ityp, tau, zv,atm

Change line 262 by

  write(io,'(a6,3f20.14)') atm(ityp(na)), (alat*tau(j,na),j=1,3)



In this way I have the atoms identified and I can replace the symbols for 
the atomic number before running CASINO.



This is not an important problem, only for the large systems with different 
atoms that I am running it is tedious replace. However, there might be other 
routines or post-processing tools that use zmesh.



Please let me know if there is another way to fix this.



Hopping to be a little usefull



Fernando Reboredo




More information about the Pw_forum mailing list