Hi, I'm not sure what you are going to calculate and if you are sure that your structure is not periodic along a specific axis (z as per the image), you can take sufficient vacuum space i.e to increase the magnitude of that specific vector. It would no longer be periodic in that direction.
Best, Em 25.02.2019 22:33, Mina Sedighi escreveu: > Hello, > > Thank you for your reply. I checked the structure and found that the error of > close molecules is related to applying periodic boundary condition. My system > is a crystalline MnO2 beside water molecules (attache image). This system is > not periodic in all directions. How can I remove PBC in Siesta? > > Regards, > > Mina > ------------------------- > > FROM: siesta-l-requ...@uam.es <siesta-l-requ...@uam.es> on behalf of sullah > <sul...@fisica.ufjf.br> > SENT: February 21, 2019 5:42:11 PM > TO: siesta-l@uam.es > SUBJECT: Re: [SIESTA-L] Error in Cholesky factorisation in cdiag > > Hi, > > Please check the latest geometry as the separation between some of the atoms > (at least two) is very small (atoms are too close). > > It is also necessary to check your initial geometry and in case if it is > correct then you need to figure it out what went wrong during the > optimization. > > SIESTA clearly informed you about the error (siesta: WARNING: Atoms "A" "B" > too close: rij = 0.000000 Ang) > > Best, > > Em 19.02.2019 21:30, Mina Sedighi escreveu: > > Dear Nick and Sullah, > > Thank you for your reply. My .fdf file commands are as follow: > > SystemLabel batt_str > SystemName battery > NumberOfSpecies 7 > NumberOfAtoms 4785 > %block ChemicalSpeciesLabel > 1 7 N > 2 6 C > 3 8 O > 4 1 H > 5 16 S > 6 30 Zn > 7 25 Mn > %endblock ChemicalSpeciesLabel > PAO.BasisType split > PAO.BasisSize DZP > PAO.EnergyShift 0.01 RY > MeshCutoff 500 Ry > kgrid.Cutoff 8.73 Ang # half of the smallest lattice vector > SCF.Mix Hamiltonian > SCF.Mixer.Method Pulay > MaxSCFIterations 10000 > SCF.H.Tolerance 1.d-3 ev > SCF.Mixer.Weight 0.3 > ElectronicTemperature 300 K > MD.TypeOfRun CG > MD.Steps 500 > MD.MaxDispl 0.05 Ang > MD.MaxForceTol 0.04 ev/Ang > atom from 4653 to 4785 > LatticeConstant 1 Ang > %block LatticeVectors > 17.4600000000 0.0000000000 0.0000000000 > -20.0000000000 34.6410161514 0.0000000000 > 0.0000000000 0.0000000000 116.9600000000 > %endblock LatticeVectors > SystemLabel.STRUCT_OUT true > SystemLabel.STRUCT_NEXT_ITER true > LongOutput true > WriteKpoints > WriteKbands > WriteCoorStep > WriteForces > WriteEigenvalues > WriteWaveFunctions > WriteMullikenPop 1 > WriteOrbMom > COOP.Write > SaveElectrostaticPotential > SaveTotalPotential > AtomicCoordinatesFormat Ang > %block AtomicCoordinatesAndAtomicSpecies ... > > ... > > ... > > %endblock AtomicCoordinatesAndAtomicSpecies > > I perform this simulation on SHARCNET (Compute Canada) graham cluster with > Siesta version 4.1-b3. Besides encountering the error that I mentioned in my > last email, I receive the following warning for some atoms: > > siesta: WARNING: Atoms "A" "B" too close: rij = 0.000000 Ang > > Thank you for your advice. > > Best regards, > > Mina > ------------------------- > > FROM: siesta-l-requ...@uam.es <siesta-l-requ...@uam.es> on behalf of Nick > Papior <nickpap...@gmail.com> > SENT: February 19, 2019 7:38:52 AM > TO: siesta-l > SUBJECT: Re: [SIESTA-L] Error in Cholesky factorisation in cdiag > > This may happen due to two atoms getting too close. > Could you check your atomic coordinates at the given MD step? > > Den man. 18. feb. 2019 kl. 22.12 skrev Mina Sedighi <msedi...@uwaterloo.ca>: > > Dear Siesta users, > > I perform a parallel relaxation simulation with siesta version 4.1-b3 on > cluster. I face the following error during simulation: > > Error in Cholesky factorisation in cdiag > Error in Cholesky factorisation in cdiag > Stopping Program from Node: 105 > Stopping Program from Node: 105 > -------------------------------------------------------------------------- > MPI_ABORT was invoked on rank 105 in communicator MPI COMMUNICATOR 3 CREATE > FROM 0 > with errorcode 1. > NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. > You may or may not see output from other processes, depending on > exactly when Open MPI kills them. > > Any help will be appreciated. > > Best regards, > > Mina > -- > > Kind regards Nick
