Developer group

aerba

Hi,

The keyword PRESSURE requires the keyword PREOPTGEOM to pre-optimize the structure at the desired pressure of 15 GPa.

It is recommended to increase the level of printing in the output with the PRINT keyword to better see what is happening.

Please, try with the following input:

ELASTCON
PRINT
PREOPTGEOM
PRESSURE
15
END
END

and let me know what you get.

aerba

Dear Marco,

Thank you for your very detailed analysis. We had a look in the code and can indeed confirm that this is a bug.

For "historical reasons", atomic radii for DFT integration in CRYSTAL are tabulated up to atomic number 98 (Cf). Einsteinium (99) is the first excluded element. Unfortunately, its radius can not be user-defined via the keyword RADIUS because in CRYSTAL atomic number 99 meant "I'm done!" (as you may have seen in the basis set input block which ends with a string "99 0").

So I am afraid there is no way to run a meaningful DFT calculation with CRYSTAL23 for Es. We are now working at a bug-fix that we will include in future versions of the code.

Sorry not to have better news this time around, and
Thanks for reporting this issue!

aerba

Can you try with:

NODIIS
LEVSHIFT
10 1
FMIXING
80

and let me know if it improves things?

aerba

Hi,

I could spot the mistake: there is a missing line for the exponent of the d-type shell of Mg:

dmitoli

Dear CRYSTAL community,

We would like to introduce the beta release of CRYSTALLine, a new desktop app to display and edit structures for CRYSTAL, to animate phonon modes, and to plot the properties a calculation returns. It's cross-platform: it runs on Linux, macOS and Windows. It is built on CRYSTALClear, and is free software under the GNU GPL v3.

CRYSTALLine follows in the tradition of MOLDRAW, written by Prof. Piero Ugliengo, which served this community for over twenty years: a program on your own machine for looking at a structure and editing it.

What it does

Handling of CRYSTAL outputs, .gui/.f34 and CIF files, each in a tab of its own, with crystallographic and calculation data beside it.
Automated live detection of symmetry (through spglib).
Easy structure editing through drag-and-drop: add, move, delete or change atoms, build supercells, switch between primitive and crystallographic settings.
Animation of vibrational modes of a frequency calculation, including out-of-Γ phonons with several exporting options.
Input builder for CRYSTAL (.d12) and PROPERTIES (.d3) from the structure on screen.
First Brillouin Zone visualiser and k-path selection.
Advanced plotting tools (from CRYSTALClear): electronic and phonon band structure, densities of states, IR and Raman spectra (harmonic and anharmonic), elastic surfaces and sections, equations of state, and simulated XRD patterns. Most plots have their options beside them and are redrawn as these are changed.
Visualisation of crystalline orbitals, charge density and electrostatic potential over the structure as isosurfaces or as slices along a lattice plane.

Installing

It needs Python 3.11 or newer, best in an environment of its own:

conda create --name crystal python=3.12
conda activate crystal
pip install CRYSTALLine
crystalline

The documentation has a small MgO example to try it, with a band structure and a phonon dispersion.

This is a beta

Accidents may take place. We would be grateful for reports of anything that fails or looks wrong, either in reply to this post or at github.com/crystaldevs/CRYSTALLine/issues. Suggestions for what to add next are more than welcome too.

Links

Documentation: crystaldevs.github.io/CRYSTALLine
Source code: github.com/crystaldevs/CRYSTALLine
PyPI: pypi.org/project/CRYSTALLine

Give it a try!

aerba

A possible "work-around" just came to my mind. As far as I understand, your job successfully completed the first deformation (STRAIN 1) to later fail during the next deformation (STRAIN 3) with the following error:

ERROR **** SCF **** SG_IRR ALLOCATION ERROR: ALREADY ALLOCATED

It looks like some arrays are not properly deallocated when moving from one deformation to the next. A possible work-around would be to compute each deformation in an independent calculation. You can use the DEFORM option to specify what deformations should be active. For instance with:

ELASTCON
PRINT
DEFORM
0 0 1 0 0 0
END

you would be asking to consider just the third deformation as active. With PRINT you ask for extended printing options. You will find the computed elastic constants for the third row of the elastic tensor printed in the output.

As for your calculation there are only three symmetry-independent deformations:

LIST OF ACTIVE DEFORMATIONS:
1
3
4

you could have all the elastic constants by running three calculations. One with:

ELASTCON
PRINT
DEFORM
1 0 0 0 0 0
END

a second one with

ELASTCON
PRINT
DEFORM
0 0 1 0 0 0
END

and a third with

ELASTCON
PRINT
DEFORM
0 0 0 1 0 0
END

Let me know if this helps.

aerba

Hi,

We have run your input files with CRYSTAL23 and the calculation was successful (see attached output file). Are you by any chance running an older version of the program?

aerba

Let me run some tests

aerba

Hi,

While there is no option to estimate the necessary memory, there are a couple of advanced printing options to monitor the memory used step-by-step throughout the calculation. Please, refer to the MEMOPRT and MEMOPRT2 keywords in the User's Manual.

Let me know if this helps,

aerba

Hi,

The input/output files you uploaded look fine. It is difficult for me to understand what happened.

Developer

Posts

What it does

Installing

This is a beta

Links

Member List