Question about poisson solver in cp2k

[Alex Brown]

Dear all,

I have a very naive question about poisson solver section below in cp2k. In some cp2k examples files I see this section included, but in some I don't. What does this section do in the calculations? Should I include this section or not in my input file? Does it affect results in any way? I did a test run with and without this section and the system energy changed by a small 1.9kJ/mol. I run periodic calculations and set PERIODIC XYZ in CELL section.

    &POISSON

       PERIODIC XYZ
       POISSON_SOLVER PERIODIC

    &END POISSON

Looking forward to the reply.

Best,

Alex

[Anna Hehn]

Dear Alex,

the section is specifying which kind of solver to use for the Poisson equation. PERIODIC XYZ and POISSON_SOLVER PERIODIC are the default values and suitable if you are running calculations on periodic systems. Adding the section thus not chances the result. You only have to change the default if you are switching e.g. to 2D or molecular systems. Periodicity in the CELL section needs to be specified in addition.

Best regards,

Anna 

--
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/74e33b74-ca70-4545-80f4-796841638ae5n%40googlegroups.com.

[Alex Brown]

Thanks Anna for the prompt response.

Best,

Alex

[Mostafa Abedi]

Hi.

I've got a question for this topic. Would that be necessary to specify "PERIODIC XYZ" in the CELL section in addition to the POISSON section in order to perform periodic calculations? I mean, would this settings be enough for periodic calculations:

&POISSON
PERIODIC XYZ
PSOLVER PERIODIC
&END

.

.

.

&CELL
ABC [angstrom] 23.0 23.0 23.0
&END CELL

Thanks,

Mostafa

[Krack Matthias (PSI)]

Hi Mostafa

 

PERIODIC XYZ is the default like PSOLVER PERIODIC. Thus, in your case, you could drop the &POISSON section and leave the &CELL section asis.

 

HTH

 

Matthias

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/8b382ef0-e9ed-4ab4-95ce-fee7801f5ebbn%40googlegroups.com.

[Mostafa Abedi]

Many thanks Matthias for your reply. So, basically there is no need for the "&POISSON " section and it can be like this for periodic and non-periodic calculations:

For periodic calculations:

&CELL
ABC [angstrom] 23.0 23.0 23.0
&END CELL

For non-periodic calculations:

&CELL
ABC [angstrom] 23.0 23.0 23.0

PERIODIC NONE

&END CELL

Right?

Mostafa

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/91DEC143-EEB2-4990-B3F0-32B5AEFD04BA%40psi.ch.

[Anna Hehn]

Dear Mostafa,

it has to be specified for non-periodic computations, so that you then have to put POISSON_SOLVER WAVELET / PERIODIC NONE in the POISSON section and PERIODIC NONE in the CELL section.

Best regards,

Anna 

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAHQ2YDxn4LNjkO_6RTWsiq5jTfRBv%2BO%3DQLOsjtPU_xx_DbRfRQ%40mail.gmail.com.