[CP2K-user] [CP2K:12516] space group

[Krack Matthias (PSI)]

Hi

There are different optimizer schemes (types) available in CP2k, but I assume you refer to the default one, namely "direct_cell_opt". For that choice your analysis is correct. The gradient vector contains the 3*nparticle (dx,dy,dz for atom 1, then for atom 2 and so forth) like for geo_opt runs in packed form, but in the case of a "direct_cell_opt" run, the 6 cell gradient components (stress tensor elements) are appended to the gradient vector and the optimizer, e.g. BFGS, minimises the atomic and cell gradients concurrently. The x0 vector is packed like the gradient vector. Take care that you check info for the right if/case branch.

HTH

Matthias

-----Original Message-----
From: cp... at googlegroups.com <cp... at googlegroups.com> On Behalf Of hut... at chem.uzh.ch
Sent: Mittwoch, 20. November 2019 07:40
To: cp... at googlegroups.com
Subject: Re: [CP2K:12516] space group

Hi

Unfortunately, I am not familiar with this part of the code.
I would have to read the routines and try to reproduce the algorithms used. Currently, you know more than I do.
If I find time I will have a look, but I cannot promise.

regards

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

-----cp... at googlegroups.com wrote: -----
To: "cp2k" <cp... at googlegroups.com>
From: "Pierre-André Cazade" 
Sent by: cp... at googlegroups.com
Date: 11/18/2019 12:42PM
Subject: Re: [CP2K:12504] space group

Dear Juerg,

I have been working in implementing the space group symmetry but I am a bit confused by the content of the x0 and g vectors as found in the subroutine geoopt_bfgs in the file bfgs_optimizer.F. From the various calls and manipulation of these vectors, I get they contained the 3*NATOM coordinates/forces of the system. Furthermore, they can contain the coordinates/forces for shells. There are also extra content which I assume are the unit cell parameters and the associated gradient. However, it is unclear where this extra information is located. From the various part of the code I read, I am under the impression that the cell gradient is located at the end of the g vector while the cell parameters are at the beginning of the x0 vector. I know this to be wrong due to various prints I have made and from how geoopt_bfgs works. Yet pieces of code like the one that follows (found in gopt_f77_methods.F) do not help making it clear:

IF (master == para_env%mepos) THEN ! we are on the master
               CPASSERT(ASSOCIATED(gradient))
               nparticle = force_env_get_nparticle(gopt_env%force_env)
               nsize = 3*nparticle
               CPASSERT((SIZE(gradient) == nsize+6))
               CALL pack_subsys_particles(subsys=subsys, f=gradient(1:nsize), fscale=-1.0_dp)
               CALL apply_cell_change(gopt_env, cell, gradient, update_forces=.TRUE.)
               cell_gradient => gradient(nsize+1:nsize+6)

Would you have a summary of what to find each of these arrays depending on the type of calculation/options requested?

Many thanks,
Pierre


On Friday, November 8, 2019 at 8:31:32 AM UTC, jgh wrote:Hi 
 
point group / space group constraints are not possible. 
 
regards 
 
Juerg Hutter
-------------------------------------------------------------- 
Juerg Hutter                         Phone : ++41 44 635 4491 
Institut für Chemie C                FAX   : ++41 44 635 6838 
Universität Zürich                   E-mail: h... at chem.uzh.ch 
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
--------------------------------------------------------------- 
 
-----c... at googlegroups.com wrote: -----
To: "cp2k" <c... at googlegroups.com>
From: "Pierre-André Cazade"  
Sent by: c... at googlegroups.com
Date: 11/07/2019 05:35PM
Subject: [CP2K:12477] space group 
 
Dear users, 
 
Is there a way to conserve/constraint the space group of crystal when performing a CELL_OPT in the same way we can constraint the lattice system with the keywords KEEP_SYMMETRY  and KEEP_ANGLES? 
 
Many thanks, 
Pierre   
  --
 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 c... at googlegroups.com. 
 To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/3a8579e9-226f-4a52-a366-843dc0d5422d%40googlegroups.com. 
  
  
  --
 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 cp... at googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/d1e7485b-a6ac-4678-b0cd-e7459552ff35%40googlegroups.com.
 

-- 
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 cp... at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/OF73947A20.8DAB1B9A-ONC12584B8.002496C4-C12584B8.002496C5%40lotus.uzh.ch.