cannot reproduce official QCG examples

[gosiao]

Hello,

I'm trying to reproduce CREST and QCG examples from the official website (https://crest-lab.github.io/crest-docs/page/examples), but I cannot reproduce all outputs.

For instance, QCG example 2 (https://crest-lab.github.io/crest-docs/page/examples/qcg/example_2.html) stops at a step of growing the cluster, without any error message. The last lines of output are:

  =========================================
  |   quantum cluster growth: GROW        |
  =========================================

 Solute:
       unit ellipsoid axis a,b,c     :   0.408   0.306   0.286
 Solvent:
       unit ellipsoid axis a,b,c     :   0.386   0.322   0.292

  solvent anisotropy            :     1.133
  solute anisotropy             :     1.169
  roff inner wall               :     1.388
  solute max dist               :    17.496
  solvent max dist              :     7.284
  inner unit axis               :     0.487     0.274     0.239
  inner ellipsoid/Bohr          :    14.889     8.363     7.291
  scaling factor outer ellipsoid:     1.000
  outer ellipsoid/Bohr          :    14.686    11.006    10.277

  Size       E        De       Detot      Density     Eatom    av. R  Rlast   Volume    Opt
            [Eh]    [kcal]     [kcal]     [u/Å^3]     [kcal]   [bohr] [bohr] [bohr^3]

which is only at ~15% of an expected output file.

I'm using CREST installation on an HPC cluster with intel 2021; the job is run sequentially, as in the example: crest benzoic_acid.xyz --qcg water.xyz --nsolv 3 --T 1 --ensemble --mdtime 50 --alpb water --wscal 1.0 --nofix

CREST version: 2.12

Are the examples listed on the website updated? How can I further inspect this issue? Thanks!

[cplett]

Hi, It seems that there is some kind of error arising that is somehow suppressed. Maybe you can add something like "2> error.out" at the end of the command line call. This should write the error to the "error.out" file. Moreover, I guess that this is a segmentation fault which was a problem of QCG in version 2.12. If possible, you may try the current continous-release version (https://github.com/crest-lab/crest/releases) that already contains a lot of bug fixes.

[gosiao]

Hi @cplett Thank you for a prompt reply!

I am actually redirecting it to a "log" file, but this does not produce anything more... Specifically, i'm running the test as crest benzoic_acid.xyz --qcg water.xyz --nsolv 3 --T 1 --ensemble --mdtime 50 --alpb water --wscal 1.0 --nofix >log.

I'm using version 02da018 in the test, which is the one you mentioned.

[gosiao]

Hi again, @cplett

I've just rerun the example with redirecting the stderr to a file as you suggested (somehow, i missed that while reading your suggestion yesterday...). It gave only 1 line: error while reading input coordinates.

All input xyz files (water.xyz and bacillaene.xyz) are in the directory; there are also solvent.xyz and solute.xyz files created in this directory, as well as subdirectories grow (with an empty qcg_grow.xyz file ) and qcg_tmp (with 3 subdirectories)..

I saw this error is produced by src/strucreader.f90 and looked inside - could it be an issue with paths/filenames?

Thanks a lot for your help!

[cplett]

Hi, This was a bug in version 2.12 if I remember correctly. Did you test a version build with the current source code? Otherwise, crest version 2.11.2 should still work, but requires the xtbiff.

[gosiao]

Hi @cplett

Thanks a lot for your message! I'm using version 02da018 in the test, which I believe is the most recent one on master...

[cplett]

Hi, I checked for this and it seems that the latest Continous release version shows this error. Thank you very much for the report. Maybe there is a compatibility problem with a certain compiler. Did you use a self-compiled version of the source code? If yes, it would be great if you could tell me which compiler you used so that I can fix it.

[gosiao]

Hi @cplett

Thanks for your message and for checking this out! I tested this with a compilation kindly provided by my colleague on the HPC side - there, it was done with intel 2021.2.0. Earlier, I tested this with my own compilation on my laptop using gfortran, but there i had other issues (#233).

In both cases (intel and gfortran), I also tested the code available through conda; same mistakes.

[cplett]

Hi, Thank you for the information. I will try to reproduce the problem so that I can fix this.

[github-actions[bot]]

This issue had no activity for 6 months. It will be closed in 1 week unless there is some new activity.