[CP2K:6559] Checking for a True Minima

[Iain Bethune]

Also, I’d add that if you have access to a parallel machine with enough cores, if you ran your initial optimisation on P cores you can run the vibrational analysis on some multiple of P cores setting the NPROC_REP keyword to P, and therefore run some (or even all!) or the 6N finite-difference calculations in parallel within a single CP2K job.

- Iain

--

Iain Bethune
Project Manager, EPCC

Email: ibet... at epcc.ed.ac.uk
Twitter: @IainBethune
Web: http://www2.epcc.ed.ac.uk/~ibethune
Tel/Fax: +44 (0)131 650 5201/6555
Mob: +44 (0)7598317015
Addr: 2404 JCMB, The King's Buildings, Peter Guthrie Tait Road, Edinburgh, EH9 3FD

> On 28 May 2015, at 09:23, Matt W <MattWa... at gmail.com> wrote:
> 
> But be careful! Vibrational analysis can be expensive as it is done numerically with finite differences - requires 6N calculations, where N is number of atoms in your system.
> 
> More importantly, DFT with some grid based calculations, like cp2k, is much more sensitive to numerical noise than quantum chemistry codes that is sounds like you are used to. This means you can easily get some imaginary frequencies due to bad numerics even when you are at a real minimum.
> 
> Matt
> 
> On Thursday, May 28, 2015 at 9:14:28 AM UTC+1, IBethune wrote:
> Hi Olivia, 
> 
> Yes, once you have the optimised structure, you can take the same input file, set RUN_TYPE to VIBRATIONAL_ANALYSIS and add a &VIBRATIONAL_ANALYSIS section.  You may remove the &MOTION section if you like, but I think CP2K will ignore it in any case. 
> 
> You need to make sure that you start from the optimised geometry (including cell), so you either need to edit the coords into your input file, or read from a file (see SUBSYS/TOPOLOGY/COORD_FILE_NAME and SUBSYS/CELL/CELL_FILE_NAME).  Assuming you are doing DFT, it will also be much quicker if you start the SCF optimisation for each of the single-point calculations required by the vibrational analysis finite-difference scheme from the optimised wavefunctions you obtained at the end of your CELL_OPT.  This can be done via the FORCE_EVAL/DFT/SCF/SCF_GUESS keyword set to RESTART. 
> 
> There are example inputs in the directory e.g. cp2k/tests/QS/regtest-gpw-2-1/H2-vib.inp that may help.  There are also some discussions about the different options that can be used in VIBRATIONAL_ANALYSIS in older posts to this group that may also be helpful. 
> 
> Cheers 
> 
> - Iain 
> 
> -- 
> 
> Iain Bethune 
> Project Manager, EPCC 
> 
> Email: ibe... at epcc.ed.ac.uk 
> Twitter: @IainBethune 
> Web: http://www2.epcc.ed.ac.uk/~ibethune 
> Tel/Fax: +44 (0)131 650 5201/6555 
> Mob: +44 (0)7598317015 
> Addr: 2404 JCMB, The King's Buildings, Peter Guthrie Tait Road, Edinburgh, EH9 3FD 
> 
> > On 27 May 2015, at 17:15, Olivia Lynes <liv... at gmail.com> wrote: 
> > 
> > Hi all, 
> > 
> > I'm a new user of CP2K and have started off doing direct cell optimisations on  32/64/128 water boxes. The aim is to introduce small atoms into the water boxes and look at the behaviour. 
> > 
> > All of the cell optimisations have completed, is there a way to check that the optimised geometry has reached a true minima not a maxima? 
> > 
> > Previously I would do a vibrational frequency analysis which gave the vibrational frequencies and infra-red spectra and if imaginary frequencies were found distort the geometry along the imaginary frequency and then re-optimise the structure. . 
> > 
> > Is there something similar I can do using CP2K? 
> > 
> > Is it just a case of once the cell optimisation is complete altering the input file so RUN_TYPE is VIBRATIONAL_ANALYSIS and including a VIBRATIONAL_ANALYSIS section? I'm a bit unsure of what the input section for this type of job should look like, would I need to remove the CELL_OPT section? 
> > 
> > If anyone has any suggestions of where I might begin with this that would be great. Apologies if the questions seem a bit simple I want to make sure I'm not misunderstanding things right at the beginning. 
> > 
> > Many thanks 
> > 
> > Olivia 
> > 
> > -- 
> > 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+... at googlegroups.com. 
> > To post to this group, send email to cp... at googlegroups.com. 
> > Visit this group at http://groups.google.com/group/cp2k. 
> > For more options, visit https://groups.google.com/d/optout. 
> 
> 
> -- 
> The University of Edinburgh is a charitable body, registered in 
> Scotland, with registration number SC005336. 
> 
> 
> -- 
> 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... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at http://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.


-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.