Finite temperature optimization

[Simone Ritarossi]

Dear users,

I have a problem with the use of CELL_OPT at temperatures other than 0K. Following CP2K's guide (I could not find any discussions on the subject on the web), I understood that I have to 1) set the RUN_TYPE CELL_OPT, 2) define TYPE MD in the CELL_OPT section and 3) define an MD section.  In the guide it says "  Two possible schemes are available: (1) Zero temperature optimization; (2) Finite temperature optimization  ". 
However, I do not get the optimization. .
I attach here the input and output files. I cannot understand the error in the output (I cannot find anything similar on the web): the abort message is " CPASSERT failed -  motion/gopt_f77_methods.F:307"


Thank you for your time!

[Thomas Kühne]

Dear Simone Ritarossi, 

just 2 comments from my side. 

(i) When using CELL_OPT/TYPE MD, a MD simulation is conducted to obtain the ensemble averaged stress tensor to conduct 

your cell optimization. This is not only excessively expensive (in your case a 10000 step AIMD calculation is performed for each 

cell optimization step!), but in my experience is also not effective due to the inherent fluctuations in the stress tensor, in particular 

when using this in conjunction with a quasi-Newton optimizer such as BFGS that tolerates noise very poorly. 

(ii) In the comment of your input file you write "Performs a geometry and cell optimization at the same time. The stress tensor is 

computed at every step“. If this is what you want to do simply use TYPE DIRECT_CELL_OPT instead. 

Greetings, 

Thomas Kühne

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/bd18aafe-a367-401d-a2cf-cad27241bb4cn%40googlegroups.com.
<PCH_GST_TiTe2_CELL_OPT_MD.out><PCH_GST_TiTe2_CELL_OPT_MD-MD-1.out><PCH_GST_TiTe2_CELL_OPT_MD.inp><PCH_GST_TiTe2_CELL_OPT_MD-MD-0.out><job.err>

[Simone Ritarossi]

Thank you very much for your reply. However, I think I don't quite understand: in CP2K I can do a CELL_OPT and get the relaxed system at 0K; can I instead get the relaxed cell at a temperature of 300K?

I could extract from the trajectory N configurations of the system at 300K and for each of these do a CELL_OPT, evaluating the volume behaviour. Does this make sense?
 
Thanks again for your time

Simone