[CP2K-user] [CP2K:21735] Re: MD dftd4 in version 2025.1

[Frederick Stein]

 Dear Somayeh,
Generally, there a bunch of changes between versions 7 and 2025.2 making it 
hard to tell what the exact issue is. Can you provide more details regarding
- What hardware are you running on?
- How did you compile CP2K in both cases?
- Can you provide profiles? It might already help to run CP2K in 
ENERGE_FORCE mode instead of MD to get some ideas.
Best,
Frederick


Somayeh Ahmadkhani schrieb am Dienstag, 12. August 2025 um 09:23:10 UTC+2:

> Dear CP2K developers and users,
>
> I am running CP2K for a water system containing approximately 2,600 
> molecules.
> With CP2K version 7, I was able to save around 50 steps in a 12-hour run. 
> However, after switching to a newer version, I now obtain data for only 2 
> steps in the same 12-hour period. Furthermore, when using the DFT-D4 
> correction, I can only obtain 1 step per 12 hours.
>
> Could you please help me understand and resolve this performance issue?
> I have attached my SBATCH script and input file for more details, in case 
> they are useful.
>
> Thank you in advance for your support.
>
> Best regards,
> my SBATCH script is as follows
>
> #!/bin/bash
>
> #SBATCH --job-name=cp2k
>
> #SBATCH --ntasks=64
>
> #SBATCH --mem=300G
>
> #SBATCH --nodes=4
>
> #SBATCH --time=12:00:00
>
> #SBATCH --error=cp2k_job.err
>
> module load intel/
>
> module load impi/2021.14
>
> module load cp2k/2025.1
>
> export *OMP_NUM_THREADS*=1
>
> ulimit -s unlimited
>
> mpirun  cp2k.psmp -i PBE_water.inp -o out_put.out
>

-- 
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/526f1667-fd26-4616-85c5-6c64de88eb9fn%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250812/df5903e3/attachment.htm>