[CP2K:193] Re: Memory leaks?

Teodoro Laino teodor... at gmail.com
Sat Jul 14 22:06:34 UTC 2007


Yep can confirm.. isolated right now the same subroutines  
(qs_collocate_density.F qs_integrate_potential.F and cube_utils.F) ..
I'm working on these files right now..

Teo
On 14 Jul 2007, at 23:57, jgh wrote:

>
> Sorry for coming late to the party.
>
> I run some applications on thursday/friday and encountered similar
> problems.
> I didn't realize until now that the problems were memory related. My
> jobs got
> slower and slower and finally crashed (swapping of course). I then
> tracked
> the problem to the changes made on July 3/4 to the files
> qs_collocate_density.F and qs_integrate_potential.F .
> The problem is related to the changes that include "contained"
> subroutines
> in module subroutines.
> I didn't have time to follow this further or to change the routines to
> full
> subroutines.
>
> Juerg
>
> On Jul 14, 9:57 pm, Teodoro Laino <teodor... at gmail.com> wrote:
>> Mmmmmmm.....
>> Unfortunately there are no such preprocessing Axel oriented  
>> things ;-))
>>
>> And moreover here is the news:
>> I'm not so happy to say that with a totally different input file I
>> was able to reproduce on the CRAY at CSCS the same problem
>> you're experiencing Axel..
>>
>> I ran out of memory after 2 geometry optimization steps.. :(((((
>>
>> I'm finishing now to isolate the commit that was creating the
>> trouble.. 1 afternoon job :((
>> hold on.. more news in the next hours..
>>
>> Teo
>
>
> >




More information about the CP2K-user mailing list