[CP2K-user] [CP2K:21737] MD dftd4 in version 2025.1
Frederick Stein
f.stein at hzdr.de
Tue Aug 12 09:34:16 UTC 2025
Dear Somayeh,
If D4 is significantly more expensive than D3, how have you determined the
cutoffs for D4? Have you made a convergence test on the accuracy of
energies and forces?
Best,
Frederick
Somayeh Ahmadkhani schrieb am Dienstag, 12. August 2025 um 11:20:18 UTC+2:
> Dear Frederick,
>
> Thank you for your reply. Please find the details below:
>
> -
>
> *Hardware:* *I am running on an HPC cluster with CPU model and I was
> shared my running script before.*
> -
>
> *System size:* My system contains *2454 molecules of water around
> 7362 atoms* in total.
> -
>
> *Compilation details:*
> -
>
> For CP2K *version 7*, as follows
>
> CP2K version 7.1
>
> Source code revision git:e635599
>
> cp2kflags: omp libint fftw3 libxc elpa parallel mpi3 scalapack
> xsmm spglib mkl
>
> compiler: Intel(R) Fortran Intel(R) 64 Compiler Classic for
> applications running on Intel(R) 64, Version 2021.13.0 Build 20240602_000000
>
> compiler options:
>
>
> -I/sw/chem/cp2k/2024.1/cpu-clx/intel_2024.2-impi_2021.13//spglib-2.5
>
> .0/install/include
> -I/sw/chem/cp2k/2024.1/cpu-clx/intel_2024.2-impi_
>
>
> 2021.13//elpa-new_release_2023.11.001/install/include/elpa_openmp-20
>
> 23.11.001/modules
> -I/sw/chem/cp2k/2024.1/cpu-clx/intel_2024.2-impi_2
>
>
> 021.13//elpa-new_release_2023.11.001/install/include/elpa_openmp-202
>
> 3.11.001/elpa
> -I/sw/chem/cp2k/2024.1/cpu-clx/intel_2024.2-impi_2021.
>
> 13//libxsmm-1.17/install/include
> -I/sw/chem/cp2k/7.1/clx.el9/intel_2
>
> 024.2-impi_2021.13//libxc-4.3.4/install/include
> -I/sw/chem/cp2k/2024
>
>
> .1/cpu-clx/intel_2024.2-impi_2021.13//libint-v2.6.0-cp2k-lmax-5/inst
>
> all/include -I/sw/compiler/intel/oneapi/mkl/2024.2/include/fftw/
> -I/
>
> sw/compiler/intel/oneapi/mkl/2024.2/include
> -I/sw/chem/cp2k/7.1/clx.
>
> el9/intel_2024.2-impi_2021.13/cp2k-7.1/src/start/
> -I/sw/chem/cp2k/7.
>
>
> 1/clx.el9/intel_2024.2-impi_2021.13/cp2k-7.1/obj/local-static/psmp/e
>
> xts/dbcsr -I/sw/comm/impi/mpi/2021.13/include/mpi
> -I/sw/comm/impi/mp
>
> i/2021.13/include -I/sw/comm/impi/mpi/2021.13/include/mpi -c -O2
> -fP
>
> IC -fp-model precise -funroll-loops -qopenmp -qopenmp-simd
> -xCORE-AV
>
> X512 -m64 -D__parallel -D__MPI_F08 -D__SPGLIB -D__ELPA
> -D__LIBXSMM -
>
> D__LIBXC -D__LIBINT -D__FFTW3 -D__FFTW3_MKL -D__MKL
> -D__SCALAPACK -d
>
> iag-disable=8291 -diag-disable=8293 -diag-disable=10448 -fpp
> -fpscom
>
> p logicals -free -std08 -D__COMPILE_ARCH="local-static"
> "-D__COMPILE
>
> _DATE="Fri Sep 27 15:48:17 CEST 2024""
> -D__COMPILE_HOST="bcn1357" -D
>
> __COMPILE_REVISION="git:e635599"
> -D__DATA_DIR="/sw/chem/cp2k/7.1/clx
>
> .el9/intel_2024.2-impi_2021.13/cp2k-7.1/data"
> -D__SHORT_FILE__="star
>
> t/cp2k.F"
> -
>
> For CP2K *version 2025.1*,
>
> CP2K version 2025.1
>
> Source code revision git:9635df4
>
> cp2kflags: omp libint fftw3 fftw3_mkl libxc elpa parallel
> scalapack mpi_f08 cosma xsmm spglib mkl libdftd4 sirius hdf5
>
> compiler: Intel(R) Fortran Intel(R) 64 Compiler Classic for
> applications running on Intel(R) 64, Version 2021.13.1 Build 20240703_000000
>
> compiler options:
>
>
> -I/home/bzfbpela/.local/easybuild/software/sirius/7.6.1-clx.el9.inte
>
> l//include/sirius
> -I/home/bzfbpela/.local/easybuild/software/pugixml
>
> /1.15-clx.el9.intel//include
> -I/home/bzfbpela/.local/easybuild/softw
>
> are/spla/1.6.1-clx.el9.intel//include
> -I/home/bzfbpela/.local/easybu
>
> ild/software/SpFFT/1.1.0-clx.el9.intel//include
> -I/sw/dataformats/hd
>
> f5/1.14.4.3/clx.el9/intel/include
> -I/sw/numerics/gsl/2.8/x86_64.el9/
>
> /include
> -I/home/bzfbpela/.local/easybuild/software/spglib/2.5.0-clx
>
> .el9.intel//include
> -I/home/bzfbpela/.local/easybuild/software/elpa/
>
>
> 2024.05.001-clx.el9.intel-ifort//include/elpa_openmp-2024.05.001/mod
>
> ules
> -I/home/bzfbpela/.local/easybuild/software/elpa/2024.05.001-clx
>
> .el9.intel-ifort//include/elpa_openmp-2024.05.001/elpa
> -I/home/bzfbp
>
>
> ela/.local/easybuild/software/libxsmm/1.17-clx.el9.intel//include -I
>
>
> /home/bzfbpela/.local/easybuild/software/libxc/6.2.2-clx.el9.intel//
>
> include
> -I/home/bzfbpela/.local/easybuild/software/libint-CP2K/2.6.0
>
> -clx.el9.intel//include
> -I/home/bzfbpela/.local/easybuild/software/C
>
> OSMA/2.6.6-clx.el9.intel//include/cosma
> -I/home/bzfbpela/.local/easy
>
> build/software/dftd4/3.7.0-clx.el9.intel//include/
> -I/home/bzfbpela/
>
>
> .local/easybuild/software/dftd4/3.7.0-clx.el9.intel//include//mctc-l
>
> ib/modules/
> -I/home/bzfbpela/.local/easybuild/software/dftd4/3.7.0-c
>
> lx.el9.intel//include//dftd4/Intel-2021.1.0.20240703/
> -I/sw/compiler
>
> /intel/oneapi/mkl/2024.2/include/fftw/
> -I/sw/compiler/intel/oneapi/m
>
> kl/2024.2/include
> -I/home/bzfbpela/.local/easybuild/build/CP2K/2025.
>
> 1/system-system-clx.el9.intel-ifort/cp2k-2025.1/src/start/
> -I/home/b
>
> zfbpela/.local/easybuild/build/CP2K/2025.1/
> system-system-clx.el9.int
>
>
> el-ifort/cp2k-2025.1/obj/CP2K-2025.1-clx.el9.intel-ifort/psmp/exts/d
>
> bcsr -I/sw/comm/impi/mpi/2021.14/include/mpi
> -I/sw/comm/impi/mpi/202
>
> 1.14/include -I/sw/comm/impi/mpi/2021.14/include/mpi -c -O2
> -fPIC -f
>
> p-model=precise -funroll-loops -qopenmp -qopenmp-simd
> -xCORE-AVX512
>
> -m64 -traceback -D__parallel -D__MPI_F08 -D__INTEL_COMPILER
> -D__DATA
>
> _DIR= -D__SIRIUS -D__SPLA -D__SPFFT -D__HDF5 -D__GSL -D__SPGLIB
> -D__
>
> ELPA -D__LIBXSMM -D__LIBXC -D__LIBINT -D__COSMA -D__DFTD4
> -D__FFTW3
>
> -D__FFTW3_MKL -D__MKL -D__SCALAPACK -diag-disable=8291
> -diag-disable
>
> =8293 -diag-disable=10448 -fpp -fpscomp logicals -free -std08
> -D__CO
>
> MPILE_ARCH="CP2K-2025.1-clx.el9.intel-ifort"
> "-D__COMPILE_DATE="Mon
>
> Jun 16 17:51:46 CEST 2025"" -D__COMPILE_HOST="blogin7"
> -D__COMPILE_R
>
> EVISION="git:9635df4"
> -D__DATA_DIR="/home/bzfbpela/.local/easybuild/
>
>
> build/CP2K/2025.1/system-system-clx.el9.intel-ifort/cp2k-2025.1/data
>
> " -D__SHORT_FILE__="start/cp2k.F"
>
>
> -
>
> *Profiles:* I will prepare and share profiling results shortly. In the
> meantime, I will run CP2K in *ENERGY_FORCE*mode, as you suggested, to
> gather more information.
>
> I was also shared input file earlier.
>
> Best regards,
> Somayeh Ahmadkhani
>
> On 12 Aug 2025, at 10:34 AM, Frederick Stein <f.s... at hzdr.de> wrote:
>
> 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 many atoms (or molecules etc.) does you system contain? My first
> guess considering your box size size and your files names is ca. 4096 H2O
> molecules.
>
> - 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
>
>
--
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/b9b9a3b2-2bc3-41e2-a16d-fdc10e8c1893n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20250812/f48293d0/attachment-0001.htm>
More information about the CP2K-user
mailing list