[CP2K-user] [CP2K:21453] Re: gCP Error

Frederick Stein f.stein at hzdr.de
Thu May 8 11:30:57 UTC 2025


Dear Falk,
This is indeed a bug. The feature is not covered by our regtest suite. If 
you compiled CP2K yourself, it might help to remove or comment out (Prepend 
`!`) the allocate statement in line 150 in src/qs_gcp_method.F . I am 
currently investigating whether that actually solves the problem.
Best,
Frederick

FS schrieb am Donnerstag, 8. Mai 2025 um 13:16:36 UTC+2:

> Hi all,
>
> I'm trying to run some calculations with r2scan-D4 with TZVP-MOLOPT basis 
> sets and run into crashes when attempting to add Grimmes gCP correction. It 
> appears as if fortran runs into issues with overlap in attribution of 
> variables (already used) when the gCP block is worked on.
>
> The output stops at this section:
>
>   Pair potential geometrical counterpoise (gCP) calculation
>
>                Gloabal Parameters:     sigma =                           
>  1.0000
>                                        alpha =                           
>  0.9410
>                                        beta  =                           
>  1.4636
>                                        eta   =                           
>  1.3249
>
>                                kind          nvirt          Emiss         
>   asto
>                                   1           14.0         0.0800         
> 2.8404
>                                   2           26.5         0.3000         
> 2.7679
>                                   3           29.5         0.3000         
> 3.1207
>                                   4            5.5         0.0270         
> 1.5899
>
> SLURM returns many errors, including many that sound similar to:
>
> At line 247 of file 
> /hpc/gpfs2/sw/cp2k/2025.1-ompi5.0-gcc13.2/src/subsys/atomic_kind_types.F
> Fortran runtime error: Attempting to allocate already allocated variable 
> 'atom_of_kind'
>
> The test input files, and slurm error output is attached. 
>
> Can someone give pointers how to fix this or use gCP properly with cp2k?
>
> All help would be appreciated,
> FS
>
>

-- 
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 visit https://groups.google.com/d/msgid/cp2k/a1d88b9c-74a6-448f-8ff5-d909d3a0753fn%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250508/21d31f4e/attachment.htm>


More information about the CP2K-user mailing list