BLYP functional and smoothing

[Yifei Shi]

Dear cp2k users,

I'm trying to run a MD simulation with NPT_I ensemble of liquid water with BLYP XC functional and dispersion. Without XC functional smoothing there's a small energy drift, but the density is very close to well established values(~1.04). But when I added smoothing, although the conserved quantity is constant, the density then changed to a substantially value(~1.2). The only difference between the 2 runs is the second one has the smoothing section:
       &XC_GRID
         XC_SMOOTH_RHO  NN10
         XC_DERIV  SPLINE2_SMOOTH
       &END XC_GRID

I'm wondering if the smoothing I used was inappropriate. I also attached the input file with smoothing. Thanks very much for any suggestions!

Best,
Yifei

[hut...@chem.uzh.ch]

Hi

a few comments, but unfortunately no real solution to your problem

- BLYP is not easy to converge on a regular grid
- Smoothing works fine for constant volume but requires very high cutoffs
for pressure calculations

- I'm not aware that MD with ALMO is possible

regards

Juerg
--------------------------------------------------------------
Juerg Hutter                         Phone : ++41 44 635 4491
Institut für Chemie C                FAX   : ++41 44 635 6838
Universität Zürich                   E-mail: hut...@chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----To: cp2k <cp...@googlegroups.com>
From: Yifei Shi
Sent by: cp...@googlegroups.com
Date: 05/17/2017 10:54PM
Subject: [CP2K:9014] BLYP functional and smoothing

--

You received this message because you are subscribed to the Google Groups "cp2k" group.

To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.

To post to this group, send email to cp...@googlegroups.com.

Visit this group at https://groups.google.com/group/cp2k.

For more options, visit https://groups.google.com/d/optout.



[attachment "BLYP_smoothing.inp" removed by Jürg Hutter/at/UZH]

[Yifei Shi]

Hi Juerg,

Thanks very much for the information.
I also did NVT calculation which didn't seem to have such problem.
I forgot that ALMO MD was our local development with Rustam Khaliullin and is not updated to the trunk yet.

Best,
Yifei