Re: [CP2K:22198] Abnormal force when use r2SCAN functinal to calculate single point

[Jürg Hutter]

Hi

Your problem is related to your definition of the XC functional:
&XC_FUNCTIONAL PBE
&MGGA_X_R2SCAN
&END MGGA_X_R2SCAN
&MGGA_C_R2SCAN
&END MGGA_C_R2SCAN
&END XC_FUNCTIONAL

This uses the shortcut for the PBE functional and adds r2SCAN to it. So in total you
are using two functionals at the same time. Delete PBE from the XC_FUNCTIONAL line.
BTW you can get this also from the functional report in the output.

In addition, I would also delete the XC_GRID section. I wouldn't trust it, especially for meta functionals.

regards
JH


________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of liu wei <whispe...@gmail.com>
Sent: Friday, April 24, 2026 8:02 AM
To: cp2k
Subject: [CP2K:22198] Abnormal force when use r2SCAN functinal to calculate single point

Dear developers and users,

I use the CP2K to calculate the single point in the DP-GEN activate learning process.
Detail information are as shown blew:
r2SCAN functional, TZVP-MOLOPT-SCAN-GTH basis set, GTH-SCAN potential,
CUTOFF: 1000 ray, REL_CUTOFF: 50 ray.
But it seem the force is abnormal high.
When I use the same parameters to calculate by vasp, the force is behaving normally.
BTW., I also use the same parameters to run AIMD by CP2K, the force is also normal.
So what happened? Converges to an unstable wave function?
The attachment is input file and output file.

Looking forward to your reply!

=======================================================
CP2K (单位: eV/Angstrom, transformed from hartree/bohr)
=======================================================
原子数: 298
力的模 最大值: 19.573561 (原子 2, fx=-18.041847, fy=+0.528236, fz=+7.572121)
力的模 最小值: 0.784510 (原子 293, fx=+0.312349, fy=+0.530290, fz=-0.486503)
力的模 平均值: 10.063743

=======================================================
VASP (单位: eV/Angstrom)
=======================================================
原子数: 298
力的模 最大值: 3.659353 (原子 223, fx=-2.162021, fy=-0.971786, fz=+2.787859)
力的模 最小值: 0.126453 (原子 128, fx=+0.123854, fy=-0.016994, fz=-0.019022)
力的模 平均值: 1.091464

--
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 visit https://groups.google.com/d/msgid/cp2k/56d46289-7486-4477-ad40-5d514c600940n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/56d46289-7486-4477-ad40-5d514c600940n%40googlegroups.com?utm_medium=email&utm_source=footer>.