Re: [SIESTA-L] using files of a previous run
Dear Sophia, As far as I know you cannot use a previous DM file if you change lattice vectors. You can only re-use it if lattice vectors remain the same. Regards, Natalia From: Siesta, Self-Consistent DFT LCAO program, http://www.uam.es/siesta [siest...@uam.es] On Behalf Of Sophia Nishad [sophia.nis...@gmail.com] Sent: 22 May 2009 17:23 To: SIESTA-L@listserv.uam.es Subject: Re: [SIESTA-L] using files of a previous run Hi Heiko, Thank you very much for the quick response. I want to apply 1% strain in x-direction and relax in y and z directions. The structure is non-periodic in y, z directions (for which I have vacuum in those two directions). I know about the UseSave... option. I wanted to know in what situations a DM file from previous run can be used as an initial approximation to a new calculation. Sophia On Fri, May 22, 2009 at 10:56 AM, Heiko Dumlich mailto:heiko.duml...@fu-berlin.de>> wrote: Dear Sophia, which lattice parameter/s did you change ? This is important, because depending of the parameter you change it might make sense to reuse the old DM file or not reuse it. Also it matters how much you change it. After a while you will get a feeling, when it does make sense. The new calculation you start with the new parameter will have to converge with the new parameters you inserted, but this can be quite fast, depending on the changed parameters. Other possible files you can use are flag with UseSave... . Set it to true and have the file of the old calculation in your folder for the new run and the files will be read. You can use the .CG file, which saves the CG steps of a relaxation run. (doesn't improve relaxation runs too much from my experiences) The .XV or .ZM files for the atom positions. (mainly to save copy and paste work after relaxation runs) The .DM file for the density matrix. (Improves convergence by the most) And the .LWF file for the localized wave functions. To have "as much as possible information read" use the UseSaveData option. Best Regards Heiko -- Sophia
Re: [SIESTA-L] using files of a previous run
Hi Heiko, Thank you very much for the quick response. I want to apply 1% strain in x-direction and relax in y and z directions. The structure is non-periodic in y, z directions (for which I have vacuum in those two directions). I know about the UseSave... option. I wanted to know in what situations a DM file from previous run can be used as an initial approximation to a new calculation. Sophia On Fri, May 22, 2009 at 10:56 AM, Heiko Dumlich wrote: > Dear Sophia, > > which lattice parameter/s did you change ? This is important, because > depending of the parameter you change it might make sense to reuse the old > DM file or not reuse it. Also it matters how much you change it. After a > while you will get a feeling, when it does make sense. > > The new calculation you start with the new parameter will have to converge > with the new parameters you inserted, but this can be quite fast, depending > on the changed parameters. > > Other possible files you can use are flag with UseSave... . Set it to true > and have the file of the old calculation in your folder for the new run and > the files will be read. > > You can use the .CG file, which saves the CG steps of a relaxation run. > (doesn't improve relaxation runs too much from my experiences) > > The .XV or .ZM files for the atom positions. (mainly to save copy and paste > work after relaxation runs) > > The .DM file for the density matrix. (Improves convergence by the most) > > And the .LWF file for the localized wave functions. > > To have "as much as possible information read" use the UseSaveData option. > > Best Regards > > Heiko > -- Sophia
Re: [SIESTA-L] using files of a previous run
Dear Sophia, which lattice parameter/s did you change ? This is important, because depending of the parameter you change it might make sense to reuse the old DM file or not reuse it. Also it matters how much you change it. After a while you will get a feeling, when it does make sense. The new calculation you start with the new parameter will have to converge with the new parameters you inserted, but this can be quite fast, depending on the changed parameters. Other possible files you can use are flag with UseSave... . Set it to true and have the file of the old calculation in your folder for the new run and the files will be read. You can use the .CG file, which saves the CG steps of a relaxation run. (doesn't improve relaxation runs too much from my experiences) The .XV or .ZM files for the atom positions. (mainly to save copy and paste work after relaxation runs) The .DM file for the density matrix. (Improves convergence by the most) And the .LWF file for the localized wave functions. To have "as much as possible information read" use the UseSaveData option. Best Regards Heiko
[SIESTA-L] using files of a previous run
Dear Siesta users, I have converged results for a big system of 300 atoms. Now if I change one of the lattice parameters, and do a new calculation, what files of the previous run I can use safely? If I use the previous DM file for the new calculation, the new calculation will start from a well-converged DM, right?? Are there any other files that can be used to expedite the calculations?? Thanks in advance, -- Sophia
