[CP2K-user] [CP2K:18207] How to set multigrid cutoff for GAPW?

[Jürg Hutter]

Hi

the maxiumum exponent included in the soft basis is:

zet = gauss_exponent(lshell, rc_hard, eps_fit, 1.0_dp)

where rc_hard is the "hard" radius of the atom "HARD_EXP_RADIUS".

It is the exponent that leads to a functional value of eps_fit at a distance of
rc_hard from the atom.

regards
JH

________________________________________
From: cp2k at googlegroups.com <cp2k at googlegroups.com> on behalf of Nicholas Winner <nwinner at berkeley.edu>
Sent: Wednesday, December 14, 2022 9:10 PM
To: cp2k
Subject: Re: [CP2K:18205] How to set multigrid cutoff for GAPW?

Thanks Jurg and Marcella. I ran a few tests and it looks like with the defaults all the exponents are pretty small (e.g. <5).

How does EPSFIT used to determine the exponent cutoff? I read through some cp2k tutorial slides and all the ones I found say that the exponents are mapped, but not what the criteria exactly is.
On Tuesday, December 13, 2022 at 11:51:28 PM UTC-8 jgh wrote:
Hi

the soft basis set (printed out with MEDIUM) is used for the GPW grid part.
Typically, the max exponent of the soft basis is in the range of 5-8.

regards
JH

________________________________________
From: cp... at googlegroups.com <cp... at googlegroups.com> on behalf of Nicholas Winner <nwi... at berkeley.edu>
Sent: Tuesday, December 13, 2022 11:47 PM
To: cp2k
Subject: [CP2K:18191] How to set multigrid cutoff for GAPW?

I know there is a well defined way for setting CUTOFF in the multigrid for GPW. Following Lippert's paper (https://link.springer.com/article/10.1007/s002140050523), we can take the relative cutoff and multiply it by the maximum exponent in the basis set.

That seems like the wrong strategy for the all electron basis sets, which can have exponents over 10,000.

Is there a scheme for setting the cutoff in GAPW, and is there a reference like Lippert's paper that can help understand that scheme?

Thanks,
Nick



--
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<mailto:cp2k+uns... at googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/cfbee009-eff0-4db2-9f31-630f1242958cn%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/cfbee009-eff0-4db2-9f31-630f1242958cn%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
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+unsubscribe at googlegroups.com<mailto:cp2k+unsubscribe at googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/5b86b3b3-4c84-4b5c-9334-68015840f362n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/5b86b3b3-4c84-4b5c-9334-68015840f362n%40googlegroups.com?utm_medium=email&utm_source=footer>.

-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759EE061F3BE7681A70538A9FE19%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM.