[Pw_forum] real space non-local pseudopotential projectors
Nicola Marzari
marzari at MIT.EDU
Tue Feb 21 21:30:23 CET 2006
Dear Nichols,
We don't really have much experience with PWSCF and very large systems,
but 1000 atoms with CP (inside quantum-espresso) seems doable (of
course, it depends on how many electrons you have).
Note that the CP steps mentioned below are damped dynamics -
CVS CP can now do the ground state relaxation with conjugate-gradients,
as well, that is more efficient (in practice, even if in theory they are
comparable).
------
On 8 dual-xeon motherboards, 3.4 GHz (16 cpu, although the second CPU on
each motherboard is close to useless):
PWSCF
(10,10) CNT - (49bohr*49*14 , 125 atoms, 3 k-points)
electronic relax - 3 hours
ionic relax - it really depends on the initial configuration. Between
5 and 10 days.
CP
(10,10) CNT - (43bohr*43*37, 325 atoms)
electronic relax - 17 hours (310 damped dynamics steps)
--------
On 4 PIV 3.2 GHz
CP
(5,5) CNT - (36bohr*36*56, 248 atoms, )
electronic relax - 44 hours (260 damped dyanmics steps)
On the same machine (4 PIV, 16GB) a spin-unpolarized
protein, with ~340 atoms, took ~10 minutes per CP step.
--------
Let us know what you get !
Best,
nicola
Nichols A. Romero wrote:
> Hi,
>
> Does PWSCF use real space non-local pseudopotential projectors using
> the method of King-Smith et. al?
>
> What are the largest system that people have treated using PWSCF?
> Given excellent computer resources, could it handle 1000 atom systems
> (lighter elements)?
>
> Thanks,
> --
> Nichols A. Romero, Ph.D.
> 1613 Denise Dr. Apt. D
> Forest Hill, MD 21050
> 443-567-8328 (C)
> 410-306-0709 (O)
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
--
---------------------------------------------------------------------
Prof Nicola Marzari Department of Materials Science and Engineering
13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA
tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu
More information about the users
mailing list