Troubleshooting: changing periodicity and index to radix array not found

[Michela Cavalieri]

Hello everyone!

I have tried to troubleshoot this on my own for the past two weeks with no success. Could I get some advice? Here is my process:

1. I prepared a 256 Ga atom slab with XYZ periodicity.
   The stress tensor and pressure fluctuations matched my calculated fluctuation values. I got good convergence for temperature.
   1. I creating a bulk structure of my desired final dimensions. Vacuum padding was 1/2 of the interatomic distance of Ga-Ga bond (ca 1.35 A). GEO_OPT.
   2. NVT MD
   3. NPT MD
   4. Added vacuum layer in the z-direction to form a surface.
   5. NPT MD. (last restart, stress trajectory and output: Ga1_NPT_vac_15-1.restart Ga1_NPT_vac_15-1.stress 18202783.out)
2. I prepared an explicit solvent structure using Maestro Schrodinger and added it on top. (Ga_solvent.xyz). Total 1690 atoms. 
   I have been trying since to run GEO_OPT with XY periodicity to no avail (geoopt (2).in). All of my attempts result in the Index to Radix Array Not Found error.
   See 18821106.out for example.

So far, I have tried to:

• Reduce CUTOFF and REL_CUTOFF from 600, 60 to 400, 40.
• Rounding off my unit cell dimensions.
• Changed FFT library.
• Added vacuum in Z.
• Switched POISSON_SOLVER from WAVELET to MT.

And I am still getting the same error. I have hit a wall with my understanding of the software and I am hoping someone will help me understand what is going on. Is my only option to do XYZ periodicity? 

Thank you very much,

Michela Cavalieri

[Stein, Dr. Frederick]

Dear Michela,
Sorry for the long time to respond. I guess it is just a very special
question.
Your system may be too large for the internal FFT library. Please consider
PREFERRED_FFT_LIBRARY FFTW3. I have just run the calculation with FFTW3 but
replaced your system with hydrogen such that the underlying FFT-grids should
be the same. If you still observe an error, could you increase the print
level and send a suitable output file.
Best,
Frederick

[Michela Cavalieri]

Dr. Stein, thank you so much for your help! Sorry for the long time to respond on my end as well. For some reason, I did not receive an email notification! I am trying FFTW3 now and increased the print level to HIGH. I will update you as soon as possible.

Thank you very much!
Michela

[Michela Cavalieri]

Good morning Dr. Stein and CP2K community!

I have an update. With the library you suggested, the job ran (thank you!!). However, no .restart file was generated, so I am a little confused.

The changes I made, as mentioned earlier, include the new library you suggested + PRINT_LEVEL HIGH.

The files that were generated include:
- .out and .err (attached)
- a number of localLog.out files
- a number of nonbonded_nl.out files
- ELF, efield, hartree, ELECTRON_DENSITY, TOTAL_DENSITY, WFN cubes
- a restart.wfn file

but no restart!!! Now, my geo_opt is not done, so I was planning on restarting and I would normally write the file myself, but coordinates, cell dimensions, velocities etc were not printed as individual files or as part of the output. Is there anything I should fix in my original script to ensure a restart gets printed? I could maybe just switch back to PRINT_LEVEL LOW like I had before.

Notably for the output and error files,

1. The .err reports an OpenFabrics error.

2. The .out reports some problems, such as: "Stress Tensor not tested for 2D Systems" (aborted after); "Berry phase moments for non uniform MOs' occupation numbers not implemented".

Kindly let me know if I should upload anything else!

Thank you, 

Michela

[Frederick Stein]

Dear Michela,

These are great news! You may try to reduce the print level again to have a less verbose CP2K. I guess that the missing restart file is related to CP2K aborting before finishing the first optimization step. You will need to employ PERIODIC XYZ instead of PERIODIC XY. I am not sure which Poisson solver implements stress tensors with PERIODIC XY.

Best,

Frederick

[Michela Cavalieri]

Hello Dr. Stein, thank you for your reply! I am actually a little confused about this comment you made: " I am not sure which Poisson solver implements stress tensors with PERIODIC XY. ". I thought MT, IMPLICIT and WAVELET solvers were available for 2D systems. Does this only refer to the electrostatics? In other words, could I use PERIODIC XY unit cell (CELL) but PERIODIC XYZ electrostatics in POISSON_SOLVER?

Funnily enough, while I was trying to understand this, I found this thread from last year where the two of spoke about a similar issue! I am now trying NUMERICAL stress tensor following your advice (left the system in 2D for now).

Thank you,

[Michela Cavalieri]

Please scrap the first part of my question. I am seeing now that they should match. I will leave it at this: trying NUMERICAL stress tensor. Will update here if it works or if I get an updated error message. Thanks :)

[Frederick Stein]

No, your output file states that the given Poisson solver doesn't implement stress tensors for 2D-periodic systems. I am pretty sure that all Poisson solvers implement stress tensors for 3D-periodic systems but not necessarily for 1D or 2D systems. In a few cases, this has changed over the years.

An alternative would be to describe your system as a 3D-periodic system if the boy in the non-periodic direction carries enough vacuum.

Best

[Michela Cavalieri]

Hello Dr. Stein, it is actually not printing a restart yet despite running after I changed the STRESS_TENSOR to NUMERICAL.

Any suggestions? I think part of the problem is that I was printing the restart EACH GEO_OPT 1, and since it did not complete GEO_OPT it did not print a restart. I am now trying to run it after changing that part of the print section to EACH QS_SCF 1.

In the meantime, is there anything else that jumps out to you as a culprit?

Thank you!

Michela

---

From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Frederick Stein <f.s...@hzdr.de>
Sent: Wednesday, July 15, 2026 12:16 PM
To: cp2k <cp...@googlegroups.com>
Subject: Re: [CP2K:22326] Troubleshooting: changing periodicity and index to radix array not found

 

--
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.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/4ae1cd77-0937-4f79-9692-14aba1bac6f7n%40googlegroups.com.

[Johann Pototschnig]

Have you tried the explicit keywords? 

https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/SCF/PRINT/RESTART.html

https://manual.cp2k.org/trunk/CP2K_INPUT/MOTION/PRINT/RESTART.html

Also:

https://github.com/cp2k/cp2k/blob/master/tests/QS/regtest-kp-1/h_sym_red.inp

https://github.com/cp2k/cp2k/blob/master/tests/QS/regtest-kp-1/h_sym_red_restart.inp

[Michela Cavalieri]

Johann, thank you so much for your response. I think I have those sections already as:

&MOTION

&PRINT &RESTART SILENT ADD_LAST NUMERIC &EACH QS_SCF 1 &END EACH &END RESTART &END PRINT

(...)

and under FORCE_EVAL section:

&PRINT &RESTART ON &END RESTART &END PRINT &END SCF

I have not needed to change these specific settings in a while, and have ran multiple MD and GEO_OPT jobs in the past without this issue.
Is that what you meant? I am aware that I have RESTART SILENT in the MOTION block, but so far it has never been a problem for other MD or GEO_OPT. It would always print a restart regardless. I am now trying RESTART LOW under MOTION.

Thank you very much,

Michela

[Michela Cavalieri]

Even after changing it to this, the result did not change. No restart got printed and I am unable to continue geo_opt. Any suggestions?

&GLOBAL PROJECT "Ga_solvent_geoopt" RUN_TYPE GEO_OPT PRINT_LEVEL HIGH WALLTIME 23:30:00 SAVE_MEM T PREFERRED_FFT_LIBRARY FFTW3 EXTENDED_FFT_LENGTHS T PREFERRED_DIAG_LIBRARY ScaLAPACK &END GLOBAL &FORCE_EVAL METHOD QS STRESS_TENSOR NUMERICAL &SUBSYS &CELL ABC 30 27 70 MULTIPLE_UNIT_CELL 1 1 1 PERIODIC XY &END CELL &KIND Ga BASIS_SET DZVP-MOLOPT-SR-GTH ELEMENT Ga POTENTIAL GTH-PBE-q13 &END KIND &KIND C BASIS_SET DZVP-MOLOPT-SR-GTH ELEMENT C POTENTIAL GTH-PBE-q4 &END KIND &KIND S BASIS_SET DZVP-MOLOPT-GTH ELEMENT S POTENTIAL GTH-PBE-q6 &END KIND &KIND H BASIS_SET DZVP-MOLOPT-GTH ELEMENT H POTENTIAL GTH-PBE-q1 &END KIND &KIND Br BASIS_SET DZVP-MOLOPT-SR-GTH ELEMENT Br POTENTIAL GTH-PBE-q7 &END KIND &KIND O BASIS_SET DZVP-MOLOPT-GTH ELEMENT O POTENTIAL GTH-PBE-q6 &END KIND &TOPOLOGY COORD_FILE_NAME Ga_solvent.xyz COORD_FILE_FORMAT XYZ MULTIPLE_UNIT_CELL 1 1 1 &END TOPOLOGY &END SUBSYS &DFT BASIS_SET_FILE_NAME /anvil/projects/x-che250203/CP2K_Files/BASIS_MOLOPT POTENTIAL_FILE_NAME /anvil/projects/x-che250203/CP2K_Files/POTENTIAL &QS EPS_DEFAULT 1.0000000000000000E-10 EXTRAPOLATION USE_PREV_WF METHOD GPW &END QS &MGRID CUTOFF 400 NGRIDS 5 REL_CUTOFF 40 &END MGRID &SCF MAX_SCF 300 EPS_SCF 3.0000000000000002E-05 SCF_GUESS ATOMIC ADDED_MOS -1 &DIAGONALIZATION T ALGORITHM STANDARD &END DIAGONALIZATION &SMEAR T METHOD FERMI_DIRAC ELECTRONIC_TEMPERATURE 303.15 &END SMEAR &MIXING T METHOD BROYDEN_MIXING ALPHA 5.0000000000000002E-02 &END MIXING &PRINT &RESTART ON &END RESTART &END PRINT &END SCF &XC #should I add cutoffs? &VDW_POTENTIAL &PAIR_POTENTIAL R_CUTOFF 10 TYPE DFTD3 PARAMETER_FILE_NAME dftd3.dat REFERENCE_FUNCTIONAL "PBE" CALCULATE_C9_TERM T &END PAIR_POTENTIAL &END VDW_POTENTIAL &XC_FUNCTIONAL PBE &END XC_FUNCTIONAL &END XC &POISSON POISSON_SOLVER MT PERIODIC XY &END POISSON &END DFT &END FORCE_EVAL &MOTION &GEO_OPT TYPE MINIMIZATION MAX_DR 1.0E-03 MAX_FORCE 1.0E-03 RMS_DR 1.0E-03 RMS_FORCE 1.0E-03 MAX_ITER 200 OPTIMIZER CG &END GEO_OPT &PRINT &TRAJECTORY SILENT &EACH QS_SCF 1 &END EACH &END TRAJECTORY &VELOCITIES SILENT &EACH MD 50 &END EACH &END VELOCITIES &FORCES SILENT &EACH QS_SCF 1 &END EACH &END FORCES &STRESS SILENT &EACH QS_SCF 1 &END EACH &END STRESS &RESTART LOW ADD_LAST NUMERIC &EACH QS_SCF 1 &END EACH &END RESTART &END PRINT &END MOTION

[Michela Cavalieri]

Hello CP2K community, happy Tuesday! I am sorry to message about this issue again. Is there something obvious I am missing?

Thank you,

Michela

[Johann Pototschnig]

I tried running it with a smaller coordinate file and got Ga_solvent_geoopt-RESTART.wfn. 

For using it you need to add:

...

&DFT

     WFN_RESTART_FILE_NAME Ga_solvent_geoopt-RESTART.wfn

...

The restart files are written after finishing one SCF optimization. 

How do run it? If run on a cluster does it use a temporary directory? 

best,

Johann

[Johann Pototschnig]

Which version due you use?

 $ cp2k.psmp --version
 CP2K version 2026.2 (Development Version)
 Source code revision 170fc5735f
 cp2kflags: omp libint fftw3 libxc elpa parallel scalapack mpi_f08 cosma libxs libxsmm spglib libdftd4 dftd4_v4_2 s_dftd3 mctc-lib tblite libvori libbqb
 compiler: GCC version 13.3.0