How to set multigrid cutoff for GAPW?

82 views
Skip to first unread message

Nicholas Winner

unread,
Dec 13, 2022, 5:47:37 PM12/13/22
to cp2k
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


Marcella Iannuzzi

unread,
Dec 14, 2022, 2:17:51 AM12/14/22
to cp2k
Dear Nick

You could apply  the same scheme by considering the max exponent of the soft basis set. 
Kind regards
Marcella

Jürg Hutter

unread,
Dec 14, 2022, 2:51:28 AM12/14/22
to cp...@googlegroups.com
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...@googlegroups.com <cp...@googlegroups.com> on behalf of Nicholas Winner <nwi...@berkeley.edu>
Sent: Tuesday, December 13, 2022 11:47 PM
To: cp2k
Subject: [CP2K:18191] How to set multigrid cutoff for GAPW?
--
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<mailto:cp2k+uns...@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>.

Nicholas Winner

unread,
Dec 14, 2022, 3:10:46 PM12/14/22
to cp2k
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.

Jürg Hutter

unread,
Dec 15, 2022, 8:02:36 AM12/15/22
to cp...@googlegroups.com
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: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Nicholas Winner <nwi...@berkeley.edu>
Sent: Wednesday, December 14, 2022 9:10 PM
To: cp2k
Subject: Re: [CP2K:18205] How to set multigrid cutoff for GAPW?
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>.

Nicholas Winner

unread,
Dec 15, 2022, 12:49:40 PM12/15/22
to cp2k
Thanks Professor! I see the function in ao_utils.F and it makes sense now.
Reply all
Reply to author
Forward
0 new messages