[Pw_forum] energy-degenerate states with different irrep labels
Silvia Bakalova wrote: > Hi, > > I have one question: for spin-orbit calculations, the energy bands are > labelled with double point group notation (D3d? in my case). > > Some of the energy-degenerate states have different irrep labels and I > wonder why? > At which k-point? I think that at Gamma they have to be degenerate if time reversal symmetry holds (i.e. no magnetic field). GS > e.g. the valence band top: > > e( 45 - 46) = 7.09176 eV 2 --> G_5+ L_4+ > > e( 45 - 46) = 7.09176 eV 2 --> G_6+ L_5+ > > I would be grateful for your reply or some literature reference, as I am > not familiar with the group theory. > > Many thanks, > > Silvia > > > -- > Dr. Silvia Bakalova, > Post Doctoral Researcher, > HH Wills Physics Laboratory, > Bristol, BS8 1TL, UK > http://spectra.phy.bris.ac.uk/ > > > > > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- o o | Gabriele Sclauzero, PhD Student | | c/o: SISSA & CNR-INFM Democritos, | |via Beirut 2-4, 34014 Trieste (Italy) | | email: sclauzer at sissa.it | | phone: +39 040 3787 511 | | skype: gurlonotturno | o o
[Pw_forum] energy-degenerate states with different irrep labels
Thank you for the reply, Gabriele. Yes, this is @Gamma, the states are degenerate, but I wonder why they have different irreducible representations (G_5+ and G_6+)? Silvia Bakalova wrote: >* Hi, *>* *>* I have one question: for spin-orbit calculations, the energy bands are *>* labelled with double point group notation (D3d? in my case). *>* *>* Some of the energy-degenerate states have different irrep labels and I *>* wonder why? *>* * At which k-point? I think that at Gamma they have to be degenerate if time reversal symmetry holds (i.e. no magnetic field). GS >* e.g. the valence band top: *>* *>* e( 45 - 46) = 7.09176 eV 2 --> G_5+ L_4+ *>* *>* e( 45 - 46) = 7.09176 eV 2 --> G_6+ L_5+ *>* *>* I would be grateful for your reply or some literature reference, as I am *>* not familiar with the group theory. *>* *>* Many thanks, *>* *>* Silvia *>* *>* *>* -- *>* Dr. Silvia Bakalova, *>* Post Doctoral Researcher, *>* HH Wills Physics Laboratory, *>* Bristol, BS8 1TL, UK *>* http://spectra.phy.bris.ac.uk/ *>* *>* *>* *>* *>* *>* ___ *>* Pw_forum mailing list *>* Pw_forum at pwscf.org <http://www.democritos.it/mailman/listinfo/pw_forum> *>* http://www.democritos.it/mailman/listinfo/pw_forum * -- o o |* Gabriele Sclauzero, PhD Student | *|* c/o: SISSA & CNR-INFM Democritos, | *|*via Beirut 2-4, 34014 Trieste (Italy) | *|* email: sclauzer at sissa.it <http://www.democritos.it/mailman/listinfo/pw_forum> | *|* phone: +39 040 3787 511 | *|* skype: gurlonotturno | *o o -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/d63ab4be/attachment.htm
[Pw_forum] Relax calculation not converging
On Oct 9, 2009, at 16:13 , Duy Le wrote: > 1. Grep total energy of last few iteration, if they are converging, > you may change the electron_maxstep to 200, 300, 5000 :) not a good advice. Self-consistency should converge in less than 100 iterations (and it does, in non-pathological cases) P. --- Paolo Giannozzi, Dept of Physics, University of Udine via delle Scienze 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
[Pw_forum] Relax calculation not converging
Convergence might be slow or oscillatory, try changing mixing beta. mighfar imam JNCASR, B'lore. India. Dimpy Sharma > > Hi Quantum espresso users, > > I have been trying to perform a relax calculation in my system with 99 > atoms (semiconductor system) and I am running my calculation in 48 > processors. The calculation has finished 3 scf steps and finished 100 > iterations, but after this it stops and it is showing me the following > message . > ' convergence NOT achieved after 100 iterations: stopping' > > I have restart my calculation, still I got the same error message after > 100 iteration steps. > > Can any body give me any suggestion? > > Thanks > > Dimpy > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum >
[Pw_forum] energy-degenerate states with different irrep labels
Hi, I have one question: for spin-orbit calculations, the energy bands are labelled with double point group notation (D3d? in my case). Some of the energy-degenerate states have different irrep labels and I wonder why? e.g. the valence band top: e( 45 - 46) = 7.09176 eV 2 --> G_5+ L_4+ e( 45 - 46) = 7.09176 eV 2 --> G_6+ L_5+ I would be grateful for your reply or some literature reference, as I am not familiar with the group theory. Many thanks, Silvia -- Dr. Silvia Bakalova, Post Doctoral Researcher, HH Wills Physics Laboratory, Bristol, BS8 1TL, UK http://spectra.phy.bris.ac.uk/ -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/f4dcc61b/attachment.htm
[Pw_forum] use cvs
Hi, everyone I installed RE 4.1 in a cluster with 40 processes using compiler mpif90 and pgf90. Machine type is 86_64. OS is Linux. There is no problem during the compile. However, I received the error message when I run first example01 #Error in iotk_scan (iotk_scan.spp:717) # CVS Revision:1.7 # direction Control=2 # Error in iotk_close_read (iotk_files.spp:589) #CVS Revision:1.7 # From IOTK Library, version 1.1.0 development # Unrecoverable error (ierr=2) And more May be I can solve the problem by login CVS, I set the environment parameter: Export CVS_RSH=ssh Export CVSROOT=:pserver:anonymous at scm.qe-forge.org:/cvsroot/q-e Then cvs login I press Enter. I did not login but back to my own director. Zhifan Chen Dept of Physics Clark Atlanta University -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/40407729/attachment.htm
[Pw_forum] Relaxation of low symmetry lattices
Manoj Srivastava wrote: > Dear users, > Thank you all for quick reply. Through all the replies I am under the > impression that the forces can never be actually zero, even for perfect > FCC lattice. I dont know the exact reason yet, whether it is just > numerical or something bad on my side. So, I would like to give it one > more shot, and i have some questions on that. > > Duy Le: I will try using crystal and report to the forum. Although > relaxation stops after completing one scf cycle, so atoms actually dont > move, but still nonzero forces. > > S. Baroni: When you use cubic symmetry, do you get forces on all atoms > zero. Did you do the calculation or is this your intution? also, can one > use cubic symmetry for low symmetry unit cell as in my case? As I wrote you in my previous mail, it is not that you use cubic symmetry, but you start from cubic symmetry operations to find the actual symmetries (by removing those which do not hold anymore depending on the lattice and atomic positions). You could do the exercise to find out by hand what the symmetry should be in your case and check that the symmetry operations correspond to those found by pw.x > > which is also one order of magnitude smaller than before, so i am > guessing there might me some other parameters that are not set up > correctly and causing symmetry breaking? Would someone mind to tell me? I > dont know much about FFT grid, Isn't FFT dimension determined by the KE > cutoff? Is there anyother way of changing it? I don't know what the cause could be, anyway, since you seem to be very curious and to have some spare time for testing, I can suggest the following: in forces.f90 uncomment the section between #if defined (DEBUG) and #endif and look at the various contributions to the forces (this is present in the CVS version, I don't know if also in the released one). By looking at the single terms you might find why they do not cancel exactly. Cheers GS > > Regards, > Manoj Srivastava > University of Florida > Gainesville, FL > -- o o | Gabriele Sclauzero, PhD Student | | c/o: SISSA & CNR-INFM Democritos, | |via Beirut 2-4, 34014 Trieste (Italy) | | email: sclauzer at sissa.it | | phone: +39 040 3787 511 | | skype: gurlonotturno | o o
[Pw_forum] Relax calculation not converging
On Oct 9, 2009, at 11:56 , Dimpy Sharma wrote: > ' convergence NOT achieved after 100 iterations: stopping' > please see here: http://www.quantum-espresso.org/user_guide/ node76.html#SECTION00096020 P. --- Paolo Giannozzi, Dept of Physics, University of Udine via delle Scienze 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
[Pw_forum] Relaxation of low symmetry lattices
Dear Lorenzo and Gabriele, Thank you very much for your suggestions. I will try them out. Regards, Manoj On Fri, 9 Oct 2009, Gabriele Sclauzero wrote: > > > Manoj Srivastava wrote: > > Dear users, > > Thank you all for quick reply. Through all the replies I am under the > > impression that the forces can never be actually zero, even for perfect > > FCC lattice. I dont know the exact reason yet, whether it is just > > numerical or something bad on my side. So, I would like to give it one > > more shot, and i have some questions on that. > > > > Duy Le: I will try using crystal and report to the forum. Although > > relaxation stops after completing one scf cycle, so atoms actually dont > > move, but still nonzero forces. > > > > S. Baroni: When you use cubic symmetry, do you get forces on all atoms > > zero. Did you do the calculation or is this your intution? also, can one > > use cubic symmetry for low symmetry unit cell as in my case? > > As I wrote you in my previous mail, it is not that you use cubic symmetry, > but you start > from cubic symmetry operations to find the actual symmetries (by removing > those which do > not hold anymore depending on the lattice and atomic positions). > You could do the exercise to find out by hand what the symmetry should be in > your case and > check that the symmetry operations correspond to those found by pw.x > > > > > > which is also one order of magnitude smaller than before, so i am > > guessing there might me some other parameters that are not set up > > correctly and causing symmetry breaking? Would someone mind to tell me? I > > dont know much about FFT grid, Isn't FFT dimension determined by the KE > > cutoff? Is there anyother way of changing it? > > I don't know what the cause could be, anyway, since you seem to be very > curious and to > have some spare time for testing, I can suggest the following: in forces.f90 > uncomment the > section between #if defined (DEBUG) and #endif and look at the various > contributions to > the forces (this is present in the CVS version, I don't know if also in the > released one). > By looking at the single terms you might find why they do not cancel exactly. > > Cheers > > GS > > > > > Regards, > > Manoj Srivastava > > University of Florida > > Gainesville, FL > > > > -- > > > o o > | Gabriele Sclauzero, PhD Student | > | c/o: SISSA & CNR-INFM Democritos, | > |via Beirut 2-4, 34014 Trieste (Italy) | > | email: sclauzer at sissa.it | > | phone: +39 040 3787 511 | > | skype: gurlonotturno | > o o > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum >
[Pw_forum] Relax calculation not converging
Hi Quantum espresso users, I have been trying to perform a relax calculation in my system with 99 atoms (semiconductor system) and I am running my calculation in 48 processors. The calculation has finished 3 scf steps and finished 100 iterations, but after this it stops and it is showing me the following message . ' convergence NOT achieved after 100 iterations: stopping' I have restart my calculation, still I got the same error message after 100 iteration steps. Can any body give me any suggestion? Thanks Dimpy -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/1191e362/attachment.htm
[Pw_forum] GIPAW pseudopotentials
On Wed, October 7, 2009 17:10, Pallister, Peter wrote: > Hi all, > > I am attempting to generate a gipaw pseudopotentials, in this case for > Mg, but I am having tremendous difficulty. The wavefunctions produced > with the input below (or minute variations thereof) continuously show > discontinuities, specifically at a radius of 15bohrs. Dear Peter, the wavefunctions are generated on a logarithmic radial grid, around 15 bohr the points are very coarse, btw how are you examining the wavfunctions? (there is not file_wfcncgen='..' in your input). best regards -- Lorenzo Paulatto SISSA & DEMOCRITOS (Trieste) phone: +39 040 3787 511 skype: paulatz www: http://people.sissa.it/~paulatto/ SISSA Webmail https://webmail.sissa.it/ Powered by SquirrelMail http://www.squirrelmail.org/
[Pw_forum] Relax calculation not converging
If it is not converged, and you restart it again without any modification (or dirty trick), it should not be converged. 1. Grep total energy of last few iteration, if they are converging, you may change the electron_maxstep to 200, 300, 5000 :) or you may increase the conv_th (after getting the relaxed structure, restart the calculation with new structure and read the potential, wavefunction, lower the conv_th...) 2. Play with Mixing_mode and mixing_beta 3. Change diagonalization -- Duy Le PhD Student Department of Physics University of Central Florida. On Fri, Oct 9, 2009 at 9:59 AM, Huolin Xin wrote: > For such big unit cells, even Gamma point should work, right? > > > On Fri, Oct 9, 2009 at 7:02 AM, Ary Junior wrote: > >> Did you make any convergence tests for ecut? The set of PW that you are >> using is enough? And about your k-points sample? >> >> [] >> >> Ary Junior >> >> >> On Fri, Oct 9, 2009 at 7:38 AM, Mighfar Imam wrote: >> >>> Convergence might be slow or oscillatory, try changing mixing beta. >>> >>> mighfar imam >>> JNCASR, B'lore. >>> India. >>> >>> >>> Dimpy Sharma >>> > >>> > Hi Quantum espresso users, >>> > >>> > I have been trying to perform a relax calculation in my system with 99 >>> > atoms (semiconductor system) and I am running my calculation in 48 >>> > processors. The calculation has finished 3 scf steps and finished 100 >>> > iterations, but after this it stops and it is showing me the following >>> > message . >>> > ' convergence NOT achieved after 100 iterations: stopping' >>> > >>> > I have restart my calculation, still I got the same error message after >>> > 100 iteration steps. >>> > >>> > Can any body give me any suggestion? >>> > >>> > Thanks >>> > >>> > Dimpy >>> > ___ >>> > Pw_forum mailing list >>> > Pw_forum at pwscf.org >>> > http://www.democritos.it/mailman/listinfo/pw_forum >>> > >>> >>> ___ >>> Pw_forum mailing list >>> Pw_forum at pwscf.org >>> http://www.democritos.it/mailman/listinfo/pw_forum >>> >> >> >> >> -- >> http://lattes.cnpq.br/8221674673413336 >> >> ___ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> >> > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/bd5a92a3/attachment.htm
[Pw_forum] Relax calculation not converging
For such big unit cells, even Gamma point should work, right? On Fri, Oct 9, 2009 at 7:02 AM, Ary Junior wrote: > Did you make any convergence tests for ecut? The set of PW that you are > using is enough? And about your k-points sample? > > [] > > Ary Junior > > > On Fri, Oct 9, 2009 at 7:38 AM, Mighfar Imam wrote: > >> Convergence might be slow or oscillatory, try changing mixing beta. >> >> mighfar imam >> JNCASR, B'lore. >> India. >> >> >> Dimpy Sharma >> > >> > Hi Quantum espresso users, >> > >> > I have been trying to perform a relax calculation in my system with 99 >> > atoms (semiconductor system) and I am running my calculation in 48 >> > processors. The calculation has finished 3 scf steps and finished 100 >> > iterations, but after this it stops and it is showing me the following >> > message . >> > ' convergence NOT achieved after 100 iterations: stopping' >> > >> > I have restart my calculation, still I got the same error message after >> > 100 iteration steps. >> > >> > Can any body give me any suggestion? >> > >> > Thanks >> > >> > Dimpy >> > ___ >> > Pw_forum mailing list >> > Pw_forum at pwscf.org >> > http://www.democritos.it/mailman/listinfo/pw_forum >> > >> >> ___ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> > > > > -- > http://lattes.cnpq.br/8221674673413336 > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/d1e55813/attachment-0001.htm
[Pw_forum] Relaxation of low symmetry lattices
On Fri, October 9, 2009 09:26, Manoj Srivastava wrote: > which is one order of magnitude smaller than before :) but the KE cutoff > is ridiculously high! I also increase k points with keeping KE cutoff 50 > Ry, and that gave me - Your initial choice of k-point grid is not very wise: as you cell is much longer in the x and y direction than in the z direction you will need more points in z than in x and y. I would say about 2 - 3 times as many. e.g. you could choose: K_POINTS automatic 6 6 14 0 0 0 > Gabriele: Would you mind to tell me how can i get rid of FFT > incompatibility issue? I just told you, in my previous email. Yet I don't think it's a good idea, and the forces won't be zero anyway. Your lattice have lower symmetry than the FCC lattice (it is actually a supercell), hence you can only have zero forces up to the numerical accuracy. best regards -- Lorenzo Paulatto SISSA & DEMOCRITOS (Trieste) phone: +39 040 3787 511 skype: paulatz www: http://people.sissa.it/~paulatto/ SISSA Webmail https://webmail.sissa.it/ Powered by SquirrelMail http://www.squirrelmail.org/
[Pw_forum] Relax calculation not converging
Did you make any convergence tests for ecut? The set of PW that you are using is enough? And about your k-points sample? [] Ary Junior On Fri, Oct 9, 2009 at 7:38 AM, Mighfar Imam wrote: > Convergence might be slow or oscillatory, try changing mixing beta. > > mighfar imam > JNCASR, B'lore. > India. > > > Dimpy Sharma > > > > Hi Quantum espresso users, > > > > I have been trying to perform a relax calculation in my system with 99 > > atoms (semiconductor system) and I am running my calculation in 48 > > processors. The calculation has finished 3 scf steps and finished 100 > > iterations, but after this it stops and it is showing me the following > > message . > > ' convergence NOT achieved after 100 iterations: stopping' > > > > I have restart my calculation, still I got the same error message after > > 100 iteration steps. > > > > Can any body give me any suggestion? > > > > Thanks > > > > Dimpy > > ___ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -- http://lattes.cnpq.br/8221674673413336 -- next part -- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20091009/f543d972/attachment.htm
[Pw_forum] Relaxation of low symmetry lattices
Dear users, Thank you all for quick reply. Through all the replies I am under the impression that the forces can never be actually zero, even for perfect FCC lattice. I dont know the exact reason yet, whether it is just numerical or something bad on my side. So, I would like to give it one more shot, and i have some questions on that. Duy Le: I will try using crystal and report to the forum. Although relaxation stops after completing one scf cycle, so atoms actually dont move, but still nonzero forces. S. Baroni: When you use cubic symmetry, do you get forces on all atoms zero. Did you do the calculation or is this your intution? also, can one use cubic symmetry for low symmetry unit cell as in my case? I increased the kinetic energy cutoff to a really high value 150Ry !! forces are - atom 1 type 1 force =-0.0074 -0.03150.0310 atom 2 type 1 force = 0.0084 -0.03910.0354 atom 3 type 1 force =-0.01280.0470 -0.0401 atom 4 type 1 force = 0.01110.0044 -0.0083 atom 5 type 1 force = 0.00080.0193 -0.0181 which is one order of magnitude smaller than before :) but the KE cutoff is ridiculously high! I also increase k points with keeping KE cutoff 50 Ry, and that gave me - atom 1 type 1 force =-0.01030.0086 -0.0038 atom 2 type 1 force = 0.0535 -0.0155 -0.0025 atom 3 type 1 force = 0.06430.0039 -0.0244 atom 4 type 1 force =-0.0603 -0.01410.0310 atom 5 type 1 force =-0.04710.0171 -0.0003 which is also one order of magnitude smaller than before, so i am guessing there might me some other parameters that are not set up correctly and causing symmetry breaking? Would someone mind to tell me? I dont know much about FFT grid, Isn't FFT dimension determined by the KE cutoff? Is there anyother way of changing it? Gabriele: Would you mind to tell me how can i get rid of FFT incompatibility issue? Andrea: I changed it from e(-8) to e(-10), still no improvement. Once again, thank you all for your help. Regards, Manoj Srivastava University of Florida Gainesville, FL On Thu, 8 Oct 2009, Lorenzo Paulatto wrote: > In data 08 ottobre 2009 alle ore 10:46:13, Gabriele Sclauzero > ha scritto: > > if you use cubic it should find 2 more, but they are not compatible with > > your FFT grid (see at the beginning of the output). > > The code will only use the 2 additional operations if nr3 is a multiple of > nr1 and nr2, which is weird as it is, by default, much smaller in your > case. Anyway, if you set nr3 equal to nr1 it will use all the symmetry > operations. > > For the symmetry experts: could you please check if the code behaviour is > correct? > > cheers > > -- > Lorenzo Paulatto > SISSA & DEMOCRITOS (Trieste) > phone: +39 040 3787 511 > skype: paulatz > www: http://people.sissa.it/~paulatto/ > > *** save italian brains *** > http://saveitalianbrains.wordpress.com/ > ___ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum >
