[CPMD-list] option to avoid writing RESTART file

[Yiming Zhang]

Thank Juerg for the useful suggestion.

I found another solution is to ask CPMD not to store wavefuntion and
density, therefore to reduce the RESTART file size significantly. The trade
off is not being able to restart with old wavefunction.

Below is copied from CPMD manual:
*
STORE* * {OFF}* *[WAVEFUNCTIONS, DENSITY, POTENTIAL]*
Section: &CPMD
The *RESTART* file is *updated* every *ISTORE* steps. *ISTORE* is read from
the next line. *Default* is at the *end of the run*.
Moreover, in the same line of the number ISTORE, you can specify the number
of self-consistent iterations (with SC=number) between two updates of
RESTART file. If OFF is specified , do not store wavefunctions and/or
density (*ISTORE* is not necessary).



On 2/9/07, Juerg Hutter <hutter at pci.unizh.ch> wrote:
>
> Hi
> > Dear CPMD community,
> >
> > I am currently running a large system with 168 atoms and the RESTART
> file is
> > already 1.8G after 3 geometry opt cycles.  The time for writing the file
> is
> > even larger than the computation time.
> > Is there anyway to let CPMD not writing the RESTART file?
>
> The restart file is written after every geometry optimization step.
> However, if not requested otherwise, the file RESTART.1 is overwritten
> and the file size should be approximately constant.
>
> You can reduce the RESTART file size by using the option (see manual)
>
> COMPRESS WRITE32 (or WRITE16, WRITE8)
>
> However, wavefunctions stored with WRITE8 or WRITE16 will not be
> optimized after a restart.
>
> Finally, if you really don't want a RESTART file you can use the option
>
> BENCHMARK
>   1 0 0 0 0 0 0 0 0 0
>
> in the &CPMD section.
>
> regards
>
> Juerg Hutter
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://cpmd.org/pipermail/cpmd-list/attachments/20070209/f58e91ba/attachment.html