The ABINIT Users Mailing List ( CLOSED )

["Anglade Pierre-Matthieu" <anglade@gmail.com>]

Hi,

You may try to use intel fortran compiler and the "-check all" option on this piece of code. It will tell you (at run time) if it creates temporary copies of the data. If so it may explain your memory problem.

regards

PMA

On 2/21/07, ngalamba@cii.fc.ul.pt <ngalamba@cii.fc.ul.pt > wrote:

Dear ABINIT users

I have posted a message asking about a problem with the calculation of Hirshfeld charges using cut3d.

I have partially found the problem. It appears to exist a problem in the routine hirsh.F90 called by cut3d.F90.

The piece of code at the end of hirsh.F90 causing the error is:

******************************************

!Check on the total charge
total_zion=sum(zion(typat(1:natom)))
total_charge=sum(hcharge(1:natom))
total_weight=sum(hweight(1:natom))

******************************************

If one comments the above sums then hirsh.F90 prints out correctly the Hirshfeld atomic (in molecule) charges. The above lines make the sums of the total charges which are mostly necessary to understand if the Hirshfeld method of partitioning the charge predict neutral systems (charge division close to complete).

Anyway, just in case anyone runs into a similar problem. I will try to see why is is that this sums are causing the glibc problem, although my knowledge about it is rather poor.

Nuno Galamba

-- 
Pierre-Matthieu Anglade