npulayit and memory usage

Total energy, geometry optimization, DFT+U, spin....

Moderator: bguster

Locked
Mamikon Gulian
Posts: 20
Joined: Thu Dec 10, 2009 5:58 pm

npulayit and memory usage

Post by Mamikon Gulian » Thu Aug 12, 2010 1:53 am

Hello,

I notice that when I start a calculation that progresses slowly, for the first few SCF cycles, with each SCF cycle the amount of memory used increases. Is this due to the iteration history for the Pulay mixer getting longer (until it has 7 datasets)? Will decreasing npulayit lessen the memory load?

Thank you,
Mamikon

mverstra
Posts: 655
Joined: Wed Aug 19, 2009 12:01 pm

Re: npulayit and memory usage

Post by mverstra » Sat Sep 04, 2010 10:03 am

Hi,

it's possible that it's the pulayit - does the memory keep increasing or does it stabilize? It could also be a proper memory leak if not. I have not seen the npulayit to be an issue elsewhere, so I doubt it. Does abinit allocate much more than announced in the output file?


Matthieu
Matthieu Verstraete
University of Liege, Belgium

Locked