siesta issueshttps://gitlab.com/siesta-project/siesta/-/issues2022-07-14T06:50:38Zhttps://gitlab.com/siesta-project/siesta/-/issues/184Use byte_count in meshsubs/dhscf to estimate memory usage2022-07-14T06:50:38ZNick R. PapiorUse byte_count in meshsubs/dhscf to estimate memory usageEnd users are typically interested in knowing their calculations memory requirements.
Currently siesta only prints out the sizes of the grids.
If instead users saw the *current* memory requirements they could more accurately request ne...End users are typically interested in knowing their calculations memory requirements.
Currently siesta only prints out the sizes of the grids.
If instead users saw the *current* memory requirements they could more accurately request needed resources in their submit scripts.
My recommendation for changing the output would be:
- [ ] retain total grid size
- [ ] estimate the maximum memory allocated for any given rank
- [ ] sum the total maximum memory allocated
- [ ] sparse matrices allocated in `state_init`
- [ ] mixing history elements
This might require a subroutine that estimates this (a bit tedious), but I think that would be ok?https://gitlab.com/siesta-project/siesta/-/issues/122Util/Contrib/APostnikov files need to be updated2023-11-29T15:49:06ZAlberto GarciaUtil/Contrib/APostnikov files need to be updated**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
The Util/Contrib/APostnikov files need to be updated to the latest versions
**Description:**
<!--
Longer and de...**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
The Util/Contrib/APostnikov files need to be updated to the latest versions
**Description:**
<!--
Longer and detailed description
-->
The Util/Contrib/APostnikov files need to be updated to the latest versions in
https://gitlab.com/siesta-project/analysis-tools/visualization (and maybe also lattice-dynamics)
(Maybe we can actually remove the Contrib directory altogether)
**Supporting material:**
<!--
You might have already some material or even preliminary code
that serves to document the suggestion. Ideally, it would be in
a Gitlab project, but you could attach other (not too large) files
-->5.0.0-beta2Alberto GarciaAlberto Garciahttps://gitlab.com/siesta-project/siesta/-/issues/106DHSCF scripting the mechanisms for writing and post-processing2021-05-07T10:43:57ZNick R. PapiorDHSCF scripting the mechanisms for writing and post-processing**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
Allow better handling of sub-properties in DHSCF to allow easier IO and post-processing of certain grid-arrays.**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
Allow better handling of sub-properties in DHSCF to allow easier IO and post-processing of certain grid-arrays.Alberto GarciaAlberto Garciahttps://gitlab.com/siesta-project/siesta/-/issues/71Inclusion of spglib support for symmetrization2023-05-03T15:19:11ZAlberto GarciaInclusion of spglib support for symmetrization**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
An interface to the spglib library is now quite mature and almost ready to be merged. This issue is a placeholder...**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
An interface to the spglib library is now quite mature and almost ready to be merged. This issue is a placeholder to organize the steps to be taken, including addressing some issues before a proper merge request.
**Description:**
<!--
Longer and detailed description
-->
*Main code changes:*
Interface to [spglib](https://spglib.github.io/spglib/): `Src/spglib_f03.f90`
Module with siesta objects which then calls spglib: `m_syms.F90`. It is enough to use this module in order to access the objects symops etc...
The symmetry init routine is called from `Src/siesta_init.F`, which also outputs basic information (too verbose for the moment)
One can currently enforce symmetries on:
- structure - Src/siesta_init.F
- Forces - Src/state_analysis.F
Options for input file:
- ../Src/read_options.F90: impose_symmetries = fdf_get ('ImposeSymmetries', .true.)
- ../Src/read_options.F90: syms_prec = fdf_get ('ToleranceSymmetries', 1.d-6)
*Rapidly implementable other uses for the symmetries:*
- k-point grid reduction (implies completing the density matrix, maybe states as well, using symmetries)
- reduction of perturbations in DFPT (idem). To be discussed with Miguel eventually
- things for images and constraints that Arsalan wanted to do, with NEB or other algorithms acting on instances with different space groups.
**Supporting material:**
<!--
You might have already some material or even preliminary code
that serves to document the suggestion. Ideally, it would be in
a Gitlab project, but you could attach other (not too large) files
-->
The code is currently in [this branch](https://gitlab.com/mjv500/siesta/-/tree/spglib-inclusion)https://gitlab.com/siesta-project/siesta/-/issues/59[output]2023-02-10T13:03:25ZNick R. Papior[output]**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
Clean output information
It would be nice to clean the output of siesta.
Some thoughts:
- [ ] Divide the output ...**Summary:**
<!--
Summarize the topic of your feature request/suggestion. Try to use a
descriptive title for the issue.
-->
Clean output information
It would be nice to clean the output of siesta.
Some thoughts:
- [ ] Divide the output into 2 parts
1. [ ] Use yaml or another readable+nestable format for SCF + MD information i.e. numerical information.
2. [ ] Use regular output for information that is important for users, and optionally for developers.
- [ ] Use a `verbose` flag to determine *what* to write out, say 100 levels, default level 50
Please add to the discussion.6.0-betaJosé Mª EscartínJosé Mª Escartínhttps://gitlab.com/siesta-project/siesta/-/issues/28Angle Constraints Block2021-05-07T09:30:23ZPeter KOVALAngle Constraints BlockConstraining of angles in the constraints block does not seem to work properly.
If I am doing calculation with a cubic unit cell and trying to constrain all angles and varying the unit cell, then the relaxation goes on, but the lenght o...Constraining of angles in the constraints block does not seem to work properly.
If I am doing calculation with a cubic unit cell and trying to constrain all angles and varying the unit cell, then the relaxation goes on, but the lenght of the unit cell along z direction (third direction) does not change.
If I try to constrain only one angle (alpha), then the first unit cell vector gets fixed.
Below is the input fdf file. I am also attaching the corresponding output file.
```
SystemName siesta
SystemLabel siesta
NumberOfSpecies 1
NumberOfAtoms 4
%block ChemicalSpecieslabel
1 79 Au.gga.1
%endblock ChemicalSpecieslabel
%block PAO.BasisSizes
Au.gga.1 DZP
%endblock PAO.BasisSizes
LatticeConstant 1.0 Ang
%block LatticeVectors
4.080000000000000 0.000000000000000 0.000000000000000
0.000000000000000 4.080000000000000 0.000000000000000
0.000000000000000 0.000000000000000 4.080000000000000
%endblock LatticeVectors
AtomicCoordinatesFormat Ang
%block AtomicCoordinatesAndAtomicSpecies
0.000000000 0.000000000 0.000000000 1
0.000000000 2.040000000 2.040000000 1
2.040000000 0.000000000 2.040000000 1
2.040000000 2.040000000 0.000000000 1
%endblock AtomicCoordinatesAndAtomicSpecies
DM.NumberPulay 3
MaxSCFIterations 1000
DM.MixingWeight 0.1
DM.UseSaveDM False
%block Geometry.Constraints
cell-angle alpha
%endblock Geometry.Constraints
XC.functional GGA
XC.authors PBE
PAO.EnergyShift 0.10000000 eV
MeshCutoff 2721.1386 eV
SpinPolarized False
MD.TypeOfRun CG
MD.NumCGsteps 200
MD.MaxForceTol 0.04000000 eV/Ang
MD.VariableCell True
#KPoint grid
%block kgrid_Monkhorst_Pack
2 0 0 0.0
0 2 0 0.0
0 0 2 0.0
%endblock kgrid_Monkhorst_Pack
```
[siesta.out](/uploads/25311ab1c4cb244538201720495a8db0/siesta.out)WishlistNick R. PapiorNick R. Papior