Cannot run the relaxation for multi atoms

You must be doing something else wrong, this has basically worked for ages ;)

Also, it is not easy to see what you are actually doing since you are only pasting a subset of the fdf input.

Dear Nick Papior

I have attached the working file in the following path.
https://drive.google.com/drive/folders/1q5z-8musueI3ukx0jFKPCen0Ld-jLVi5?usp=sharing

Do you need other information?

Best regards
Jiang, Jyun-Jie

Nick Papior <email address hidden> 於 2021年2月18日 週四 下午4:55寫道：

> Your question #695616 on Siesta changed:
> https://answers.launchpad.net/siesta/+question/695616
>
> Status: Open => Needs information
>
> Nick Papior requested more information:
> You must be doing something else wrong, this has basically worked for
> ages ;)
>
> Also, it is not easy to see what you are actually doing since you are
> only pasting a subset of the fdf input.
>
> --
> To answer this request for more information, you can either reply to
> this email or enter your reply at the following page:
> https://answers.launchpad.net/siesta/+question/695616
>
> You received this question notification because you asked the question.
>

Hi,

1. Don't use 4.1-b4, use the latest stable release 4.1.5

2. You have compiled a serial Siesta version.

3. You are runring the serial Siesta in parallel (i.e. same calculation for ~ 20 cores)

4. Problem comes from double writing to the same file.

Please compile siesta in parallel, and retry, ask your local HPC admin for instructions/help.

Dear Nick Papior

I install the latest stable release 4.1.5. However, it will show the error about `GFORTRAN_1.6' not found when I run the file.
How can I solve this problem?

Dear Nick Papior

I install the latest stable release 4.1.5. However, it will show the error about `GFORTRAN_1.6' not found when I run the file.
How can I solve this problem?

Here is the working file.
https://drive.google.com/drive/folders/1mdW-V_W3QhWSL2-Z6JFPg_-8kEtf3Q5k?usp=sharing

I appreciate your help.

Best regards
Jiang, Jyun-Jie

That sounds like you are not using the same environment as you compiled siesta in.

Please ask your local IT department on how to use the same environment, and also to help correct linker lines.

Lastly, please carefully read the compilation instructions in the manual.

Dear Nick Papior

When I run the LDA+U, it shows the below error message.
Whether I do not install some files?

                           ***********************
                           * WELCOME TO SIESTA *
                           ***********************

reinit: Reading from input.fdf

reinit: -----------------------------------------------------------------------
reinit: System Name: Ga2O3
reinit: -----------------------------------------------------------------------
reinit: System Label: Ga2O3
reinit: -----------------------------------------------------------------------

initatom: Reading input for the pseudopotentials and atomic orbitals ----------
Species number: 1 Atomic number: 31 Label: Ga
Species number: 2 Atomic number: 8 Label: O

Reading PAOs and KBs from NetCDF/ascii files is not possible with LDAU.Proj
Stopping Program from Node: 0

Best regards
Jiang, Jyun-Jie

You can't supply a user-basis with LDA+U calculations.
So you have to do it from the options of siesta for these kinds of calculations.

I think the error message is pretty clear here?

Dear Nick Papior

Sorry, I cannot understand what you mean.
Whether you mean that I cannot utilize the Ga.ion and O.ion?
How can I do it from the options of Siesta?
I want to relax my structure with LDA+U. Do you have any examples?

Here is my input file.

SystemName Ga2O3
SystemLabel Ga2O3

==================================================
==================================================
# SPECIES AND BASIS

# Number of species
NumberOfSpecies 2
%block ChemicalSpeciesLabel
 1 31 Ga
 2 8 O
%endblock ChemicalSpeciesLabel

#PAO.BasisSize DZ
#PAO.EnergyShift 0.05 Ry
User.basis true

==================================================
==================================================
# K-points

%block kgrid_Monkhorst_Pack
 1 0 0 0.0
 0 1 0 0.0
 0 0 1 0.0
%endblock kgrid_Monkhorst_Pack

==================================================
==================================================
# UNIT CELL AND ATOMIC POSITIONS

# UNIT CELL
LatticeConstant 1.00 Ang
%block LatticeVectors
 12.45244598 0.000000000 0.0000000000
 0 3.082968 0.0000000000
 -1.39006327 0.000000000 5.70936982
%endblock LatticeVectors

# Atomic coordinate
NumberOfAtoms 20
AtomicCoordinatesFormat ScaledCartesian
%block AtomicCoordinatesAndAtomicSpecies
1.573998500 2.312226000 1.748009910 1
3.262174500 0.770742000 3.961358910 1
0.115212500 0.770742000 4.549389910 1
4.720959500 2.312226000 1.159977910 1
7.800212500 0.770742000 1.748003910 1
9.488406500 2.312226000 3.961366910 1
6.341424500 2.312226000 4.549387910 1
10.947193500 0.770742000 1.159985910 1
5.930821500 0.770742000 1.466002910 2
-1.094650500 2.312226000 4.243376910 2
1.457260500 0.770742000 3.229967910 2
3.378911500 2.312226000 2.479400910 2
2.636064500 2.312226000 5.147203910 2
2.200106500 0.770742000 0.562165910 2
12.157031500 2.312226000 1.465999910 2
5.131583500 0.770742000 4.243381910 2
7.683485500 2.312226000 3.229968910 2
9.605133500 0.770742000 2.479409910 2
8.862292500 0.770742000 5.147202910 2
8.426332500 2.312226000 0.562168910 2
%endblock AtomicCoordinatesAndAtomicSpecies

==================================================
==================================================
# Geometry Constraints

#%block GeometryConstraints

#%endblock GeometryConstraints

==================================================
==================================================

# General variables

ElectronicTemperature 295 K
MeshCutoff 300. Ry
xc.functional LDA # Exchange-correlation functional
xc.authors PZ
SpinPolarized .true.
SolutionMethod Diagon
#MD.MaxForceTol 0.01 eV/Ang # Tolerance in force
==================================================
==================================================
# SCF variables

DM.MixSCF1 T
MaxSCFIterations 500 # Maximum number of SCF iter
DM.MixingWeight 0.01 # New DM amount for next SCF cycle
DM.Tolerance 1.d-5 # Tolerance in maximum difference
DM.UseSaveDM true # to use continuation files
DM.NumberPulay 5
Diag.DivideAndConquer no
Diag.ParallelOverK no

==================================================
==================================================
# MD variables

MD.FinalTimeStep 1
MD.TypeOfRun CG
MD.NumCGsteps 800
MD.UseSaveXV .true.
MD.ConstantVolume .true.
UseSaveData .true.
==================================================
==================================================
# Output variables

WriteMullikenPop 1
WriteBands .false.
SaveRho .true.
SaveDeltaRho .true.
SaveHS .false.
SaveElectrostaticPotential false
SaveTotalPotential false
WriteCoorXmol .true.
WriteMDXmol .false.
WriteMDhistory .true.
WriteEigenvalues true
COOP.Write .false.

==================================================
==================================================

#PDOS calculation
#%block ProjectedDensityOfStates
#%-6.2484 -0.2484 0.04 1000 eV
#%endblock ProjectedDensityOfStates

==================================================
==================================================

LDAU.FirstIteration .false.
LDAU.ThresholdTol 1.0d-3
LDAU.PopTol 4.0d-4
LDAU.ProjectorGenerationMethod 1
LDAU.CutoffNorm 0.90

%block LDAU.proj
O 1 # number of shells of projectors
n=2 1 # n, l
4.000 0.000 # U(eV), J(eV)
0.000 0.000 # rc, \omega (default values)
%endblock LDAU.proj

==================================================
==================================================

#%BandLinesScale ReciprocalLatticeVectors

#%block BandLines
#%1 0.5 0.5 0.5 R
#%20 0 0 0 G
#%10 0.5 0 0 X
#%10 0.5 0.5 0 M
#%16 0 0 0 G
#%endblock BandLines

==================================================
==================================================

This option:

User.basis true

cannot be used with DFT+U. So the question is how you generated those ion files?
I don't know how to be more clear.

The basis set options + psf creates the ion files, so if you did this originally you should just use the basis set options used to generate your ion files directly in the fdf file...

Dear Nick Papior

1. Whether you mean I should generate .ion file again by utilizing .psf and .psml files with the fdf file having LDA+U? Then I can receive a new .ion file with LDA+U.

2. Where I can get the .psml files?
In the past, there were on the website, but I cannot find it on the website now.
(https://departments.icmab.es/leem/siesta/Pseudopotentials/index.html)

I appreciate your help.

Best regards
Jiang, Jyun-Jie

1. No, you can NEVER use ion files with DFT+U. It seems to me that you don't fully know how siesta uses basis sets and pseudo-potentials? I would highly recommend you spend some time reading the manual related to basis sets etc. The way you are doing it is cumbersome and very complicated.

2. We don't share these anymore since they are not good pseudo's for many systems. Instead we encourage users to create their own pseudopotentials that suits their chemical environment. I cannot help you with this, you need to do this. :)

Dear Sir

When I run the Siesta, it shows this error

CH_OVERLAP: THE NUMBER OF INTEGRATION POINTS MUST BE INCREASED
ch_overlap: rmx,rmax = 35.775439 32.169909
Stopping Program from Node: 0

I have read the manual and code[rmax=b*(exp(a*(nmax-1))-1.0d0)].
I tried these parameters, but the problem cannot be solved.
Reparametrize.Pseudos .false.
Rmax.Radial.Grid 40
New.A.Parameter
New.B.Parameter

Could you give me some suggestions to increase the value of rmax?
I appreciate your help.

Best regards
Jiang, Jyun-Jie

Please:

-- Start by getting familiar with the way in which Siesta works. There are a number of tutorials linked in the Documentation section of the web page https://siesta-project.org.

-- Look in the Tests directory for input files that exploit a feature similar to what you want.

-- Make sure you understand, by reading the manual, what every keyword is doing. If you cannot, go back to step 1.

-- Modify appropriately the inputs to suit your calculation, testing at every step.

-- *If* everything else fails, please post a question in the Siesta mailing list. Somebody might be able to help you. But you will need to state your problem appropriately and supply the relevant information.