The ABINIT Users Mailing List ( CLOSED )

[TORRENT Marc <marc.torrent@cea.fr>]

Hello everybody,
>   The amount of memory required by the outkss routine 
>  in case of diagonalization is estimated in memkss.F90. 
>  One can have an idea of the memory allocate by just looking for 
>  the following line in the log file : 
>
>  Additional amount of memory required by "outkss" ....
>   
I wrote this routine at the first beginning of outkss 's life. But it 
doesn't include later modifs from Valerio (vkb terms and possibility of 
using CG eigenvalues instead of exact diagonalization ones).
In principle it should be an overestimated estimation of the memory as 
the complete diago requires more memory.

>  We should write something similar also in screning and sigma.
>
>  Concerning the mkmem=0 option, it seems it's implemented in outkss 
>  but I've never used this possibility.  
>   
Same remark: this option was available at the beginning (and it worked).
But I do not see any reason why it couldn't work now.

Marc