[Pw_forum] memory

Stefano Baroni baroni at sissa.it
Sat Apr 16 13:47:53 CEST 2005


Dear Jaita,

the "mistery" should only arise for a gamma-only calculation. In this 
case the number of plane waves, strictly speaking, remains roughly the 
same, but only half of them are really used because, due to 
time-reversal invariance, wavefunctions at k=0 can always be chosen to 
be real. i.e. the number of PW's is the same, but the number of 
plane-wave coefficients (which is what matters memory-wise) is halved.

Stefano

On Apr 16, 2005, at 1:01 PM, Jaita Paul wrote:

> Dear all,
>
> Before running pw.x calcs i gave a check on the memory estimate of the
> file.with 3 k points the estimated mem was 5.2 GB for a unit cell with 
> 91
> atoms-when i reduce the k-points to a single(gamma)point then also the
> estimated mem remains the same. the no. of plane waves reduces to
> half(from 31000 to 16000) when i change the k point mesh.i cannot 
> figure
> out the mistake....
> can anyone help me?cud this arise out of wrong input atomic coords?
>
> regards,
> Jaita.
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
---
Stefano Baroni    ---  SISSA  &  DEMOCRITOS National Simulation Center
via Beirut 2-4 34014 Trieste Grignano / [+39] 040 3787 406 (tel) -528 
(fax)

Please, if possible, don't  send me MS Word or PowerPoint attachments
Why? See:  http://www.gnu.org/philosophy/no-word-attachments.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 1441 bytes
Desc: not available
Url : /pipermail/attachments/20050416/583a5f63/attachment.bin 


More information about the Pw_forum mailing list