[Wien] Comparing Total Energies
Dear Fabien, Thanks for the response. It is probably my k-mesh that is making it inaccurate to compare. David. On Tue, Jul 10, 2012 at 3:48 PM, tran at theochem.tuwien.ac.at wrote: Hi, When comparing two total energies obtained from different unit cells, one has to be always very careful. First, if you did the calculation on the small unit cell with a k-mesh (n1,n2,n3), then the calculation on the large unit cell should be done with the corresponding k-mesh (n1/m1,n2/m2,n3/m3), where m1, m2, and m3 indicate by how many times (along the lattice vectors a, b and c) is the large unit cell larger than the small one. You can have problems if If n1/m1, n2/m2 or n3/m3 is not an integer. But this not a problem if you are using very good k-meshes such that the energies are well converged. It can also happen that the orientation of the lattice vectors are completely different in the small and large unit cells. In this case some numerical noise (e.g., from the FFTs) can introduce some errors. But actually, what do you mean by I can't compare? My recommendation: use the same unit cell. F. Tran On Tue, 10 Jul 2012, David Tompsett wrote: Dear All, I have what is probably a very basic question about comparing total energies. I have been considering a system that undergoes a charge density wave transition from a high to low symmetry structure. I have been comparing the total energy of the high and low symmetry structures. The unit cell of the high symmetry structure has half as many atoms as does the low symmetry one. In my calculations it seems I can't compare 2x(total energy of high symmetry cell) with 1x(total energy of low symmetry cell). Instead I have to perform the high symmetry calculation in the same size unit cell and symmetry as for the low symmetry. Why is this the case? What is the effect of having different sized unit cells? Thank you, David Tompsett. ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120711/e6ca7ddc/attachment-0001.htm
[Wien] Comparing Total Energies
Dear All, I have what is probably a very basic question about comparing total energies. I have been considering a system that undergoes a charge density wave transition from a high to low symmetry structure. I have been comparing the total energy of the high and low symmetry structures. The unit cell of the high symmetry structure has half as many atoms as does the low symmetry one. In my calculations it seems I can't compare 2x(total energy of high symmetry cell) with 1x(total energy of low symmetry cell). Instead I have to perform the high symmetry calculation in the same size unit cell and symmetry as for the low symmetry. Why is this the case? What is the effect of having different sized unit cells? Thank you, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120710/9ae7dc58/attachment.htm
[Wien] How to make Root3xRoot3 supercell
VESTA (http://jp-minerals.org/vesta/en/) also has good utilities for letting you redefine the unit cell vectors via a matrix transformation. David. On Sat, Jun 30, 2012 at 5:59 AM, Laurence Marks L-marks at northwestern.eduwrote: Try cryscon (http://www.shapesoftware.com), perhaps also crystalmaker or similar. On Fri, Jun 29, 2012 at 3:41 PM, H M Sohail sohail.ifm at gmail.com wrote: Simple Question: How to make Root3xRoot3 supercell, since it only excepts integers.. Ge(111)1x1 to Ge(111)R3xR3 is required.. Help is required.. Regards ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- Professor Laurence Marks Department of Materials Science and Engineering Northwestern University www.numis.northwestern.edu 1-847-491-3996 Research is to see what everybody else has seen, and to think what nobody else has thought Albert Szent-Gyorgi ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120630/b6ce3125/attachment.htm
[Wien] Strange results lapwdm
Hi Peter, Thanks for the response. The option r-index = 11 is not in the UG, but is documented in Pavel Novak's technical report on the Hyperfine Field calculation so other users may strike it in future. Hopefully he will have some input. Thanks, David. On Thu, Jun 7, 2012 at 7:47 AM, Peter Blaha pblaha at theochem.tuwien.ac.atwrote: You have to check directly in the code. Without looking in the code I do not know what r-index = 11 should do ? I guess this is not a documented option, but Pavel Novak migth have put in something Am 06.06.2012 14:11, schrieb David Tompsett: Dear All, I have been attempting to use lapwdm to obtain the expectation value of the radial distribution of orbitals i.e. r. I have attempted the simplest case that I could think of, the hydrogen atom in a large cell. I used an RMT of 4.5 au so that a fraction 0.99 of the charge lies inside the muffin tin. Then after converged SCF cycles I ran x lapwdm -up with input: -9. Emin cutoff energy 1 number of atoms for which density matrix is calculated 1 1 0 index of 1st atom, number of L?s, L1 11 1 r-index, (l,s)-index From the solution of the Schroedinger equation for a Hydrogen atom, the result should be r = 1.5 au. However the result from :XOP in the output of lapwdm is 104670.12. I do not know what the units are. Also, the value of this result seems to vary strongly with the RMT. Can anyone help me understand this behaviour or if there is a problem with lapwdm? I looked through the source code, but could not follow it. Many thanks, David. __**_ Wien mailing list Wien at zeus.theochem.tuwien.ac.**at Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.**ac.at/mailman/listinfo/wienhttp://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- Peter Blaha Inst.Materials Chemistry TU Vienna Getreidemarkt 9 A-1060 Vienna Austria +43-1-5880115671 __**_ Wien mailing list Wien at zeus.theochem.tuwien.ac.**at Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.**ac.at/mailman/listinfo/wienhttp://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120607/2806a118/attachment.htm
[Wien] Strange results lapwdm
Hi Gavin, That is a good idea, with RINDEX= -11 we obtain: WEIGHTS READ Calculation of X, X=1/r**1 atom Lup dn total product Usym*Coupl after densmat 0.98974 Density matrix, imag part 0.0 :XOP 1 0 0.98974 0.0 0.98974 This looks healthier, the answer should formally be 1/r_0 if we neglect relativistic effects. So perhaps the problem is just for krad 10? Many thanks, David. On Thu, Jun 7, 2012 at 2:53 PM, Gavin Abo gsabo at crimson.ua.edu wrote: May or may not help: RINDEX = 11 = r**krad for krad 10 = :XOP 104670.12 (large number) Adding a negative sign will change the calculation: RINDEX = -11 = 1/r**krad for krad 10 = :XOP ? (small number) File of reference: $WIENROOT/SRC_lapwdm/radint.f On 6/7/2012 7:08 AM, David Tompsett wrote: Hi Peter, Thanks for the response. The option r-index = 11 is not in the UG, but is documented in Pavel Novak's technical report on the Hyperfine Field calculation so other users may strike it in future. Hopefully he will have some input. Thanks, David. On Thu, Jun 7, 2012 at 7:47 AM, Peter Blaha pblaha at theochem.tuwien.ac.atwrote: You have to check directly in the code. Without looking in the code I do not know what r-index = 11 should do ? I guess this is not a documented option, but Pavel Novak migth have put in something Am 06.06.2012 14:11, schrieb David Tompsett: Dear All, I have been attempting to use lapwdm to obtain the expectation value of the radial distribution of orbitals i.e. r. I have attempted the simplest case that I could think of, the hydrogen atom in a large cell. I used an RMT of 4.5 au so that a fraction 0.99 of the charge lies inside the muffin tin. Then after converged SCF cycles I ran x lapwdm -up with input: -9. Emin cutoff energy 1 number of atoms for which density matrix is calculated 1 1 0 index of 1st atom, number of L?s, L1 11 1 r-index, (l,s)-index From the solution of the Schroedinger equation for a Hydrogen atom, the result should be r = 1.5 au. However the result from :XOP in the output of lapwdm is 104670.12. I do not know what the units are. Also, the value of this result seems to vary strongly with the RMT. Can anyone help me understand this behaviour or if there is a problem with lapwdm? I looked through the source code, but could not follow it. Many thanks, David. ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- Peter Blaha Inst.Materials Chemistry TU Vienna Getreidemarkt 9 A-1060 Vienna Austria +43-1-5880115671 ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien ___ Wien mailing listWien at zeus.theochem.tuwien.ac.athttp://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120607/66d2c9fd/attachment.htm
[Wien] Strange results lapwdm
Dear All, I have been attempting to use lapwdm to obtain the expectation value of the radial distribution of orbitals i.e. r. I have attempted the simplest case that I could think of, the hydrogen atom in a large cell. I used an RMT of 4.5 au so that a fraction 0.99 of the charge lies inside the muffin tin. Then after converged SCF cycles I ran x lapwdm -up with input: -9. Emin cutoff energy 1 number of atoms for which density matrix is calculated 1 1 0 index of 1st atom, number of L?s, L1 11 1 r-index, (l,s)-index
[Wien] Radial expectation lapwdm
Dear All, I need a little help with the program lapwdm. I am trying to quantify the radial extent of d-orbitals of manganese ions. I wanted to use the expectation value r calculated by lapwdm using RINDEX=11, LSINDEX=1. I am not using SO-coupling and am using the latest release of Wien2k. A few questions: 1) Is RINDEX=11, LSINDEX=1 the correct setting? 2) For spin up I have :XOP 1 2 360.97696 0.0 360.97696 For spin dn I have :XOP 1 2 166.01005 0.0 166.01005 a) What units is r in? b) Also, should I divide these numbers by the occupation n_s for each spin to get the expectation? c) Even after dividing by the occupations, the values for spin up and spin down are quite different by nearly a factor of three. Seems strange to me, am I comparing the right things? 3) Is it possible to calculate the expectations for particular l,m for instance for just t_2g orbitals? Many thanks, David. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120530/ff2edf21/attachment.htm
[Wien] Questions on the quantitative understanding of DOS and partial DOS
Dear Balamurugan, Some answers below and many previous postings: On Thu, Mar 22, 2012 at 4:24 PM, J. K. Balamurugan albertbalagan at gmail.comwrote: Dear WIEN2k developers and users, I have some problem in quantitative understanding the DOS and partial DOS plots/data. I post the question with a case assumption of room temperature (RT) structure of BaTiO3. As it is well known BaTiO3 has tetragonal structure at RT. Now, Ba and Ti has one site and O has two sites. If the DOS partial DOS are calculated for BaTiO3, I will get total DOS for the unit cell which contains one formula unit of BaTiO3. I will also get partial DOS of Ba, Ti and O1 and O2 (in terms of total and from individual orbital contribution like s, p and d whichever are applicable.) Here are my questions: 1. Will the sum of total DOS individual atoms [i.e., total DOS of (Ba + Ti + O1 + O2)] be exactly equal to total DOS of the unit cell? If the answer is No, why? No, the interstitial states contribution is missing. Only the area inside the muffin tin is included in the partial DOS. 2. Will the total DOS of any specific atom, say Ti, will be exactly equal to the sum of orbital contributions [i.e., partial DOS of (Ti-s + Ti-p + Ti-d)]? If the answer is No, why? Yes, but I think even further depending on how many l-values are in the basis set. Another related question: Let us assume a hypothetical structure where in I have two formula units of BaTiO3 in a unit cell. That is the unit cell has Ba2Ti2O6. Now the question is the following: 3. Should I need to multiply 2 to the total DOS of individual atoms {i.e., 2*[total DOS of (Ba + Ti + O1 + O2)]} to get the total DOS of the unit cell with two formula unit? Will this sum be exactly equal to the actual total DOS of the unit cell which we get as it is from the calculations? Yes. Then no not total DOS of cell, see point 1. Please explain me to clear these my questions. I wish also to get into reading any material/user guide/article/document which could help me to have a complete understanding on this issue. Thanks. With kind regards, K. Balamurugan. David Tompsett. -- *K. Balamurugan Pittsburgh, USA. +1 412 961 5055* ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120322/c3070342/attachment.htm
[Wien] B_lat hyperfine field NMR
Dear All, Has anyone implemented a way to calculate the hyperfine field due to dipolar interactions between the nuclear spin and the electron spin density outside the muffin tin sphere? This is referred to as B_lat in Pavel Novak's notes on hyperfine calculations in Wien2k. Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120301/4daec80d/attachment.htm
[Wien] Suggestion for LOPW error
Dear All, There have been several discussions over the past year regarding errors of LOPW plane waves exhausted in high symmetry cells. I have recently come across the same problems in a cubic (space group 212) cell. I could only avoid them by taking RKmax up to 9.5. This however, seems to be nearing the point of overcompleteness in the basis set?
[Wien] Installation question
Dear All, I have recently attempted to recompile Wien2k version 11 and am having some problems with the MPI parallel binary lapw1_mpi. I used a GNU reference implementation of Scalapack and libsci. Unfortunately my cluster does not have inter libraries. The compile completes with no errors, but upon running lapw1_mpi does not print any eigenvalues, yet throws no errors: EIGENVALUES ARE: 0 EIGENVALUES BELOW THE ENERGY -10.0 There is simply nothing printed in case.output1up or case.energyup where the eigenvalues should be. Has anyone seen such a problem before? Also, which part of the source code should print these eigenvalues? I would like to check if the problem is with IO by trying to flush the stream. Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20120130/87d4f58f/attachment.htm
[Wien] Calculation for Heavy atom
Dear All, Is it possible to perform a calculation in Wien2k for the charge density of a very heavy atom eg. #116 or 120? Also, would the answers be sensible? Eg. for #116 I tried this case.inst: Up Rn 5 5, 3,3.0 N 5, 3,3.0 N 5,-4,4.0 N 5,-4,4.0 N 6, 2,1.0 N 6, 2,0.0 N 7,-1,1.0 N 7,-1,1.0 N 7, 1,1.0 N 7, 1,1.0 N END of input (instgen_lapw) And case.struct Title P LATTICE,NONEQUIV.ATOMS: 1 MODE OF CALC=RELA unit=ang 18.897269 18.897269 18.897269 90.00 90.00 90.00 ATOM 1: X=0. Y=0. Z=0. MULT= 1 ISPLIT= 2 Up NPT= 781 R0=0.0100 RMT=3.5000 Z:116.0 LOCAL ROT MATRIX:1.000 0.000 0.000 0.000 1.000 0.000 0.000 0.000 1.000 But I receive errors after trying to run lstart: error: Elemant120.inst not consistent with Z edit Elemant120.inst and rerun lstart afterwards or change Z in StructGen! If I try to put 5g states into case.inst (for element #120) lstart stops with a format error. Thank you, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/2007/89941621/attachment.htm
[Wien] Magnetic order with hybrids and DFT+U
Dear All, I am considering calculations of magnetic order in transition metal anion compounds. Is anyone aware of cases where DFT+U predicts the wrong magnetic order compared to experiment, but where hybrids predict correctly? Thank you, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20111005/07c7448a/attachment.htm
[Wien] Forces using Tetra
Dear All, I have a question about the forces obtained when using TETRA in metallic systems. I read in the 1994 Blochl paper that the Blochl Tetra method is not variational with respect to partial occupancies. Does this mean that the total energy from Tetra is not smoothly variational wrt. to partial occupancies? Therefore, can I trust the calculated forces in a metallic system? Can I use the total energy from Tetra for metals reliably? Many thanks for any help, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110728/e8d09646/attachment.htm
[Wien] Forces using Tetra
Hi Laurence, Is there any explanation or reference as to why they are OK? The Bloch paper suggests that they won't be. Cheers, David. 2011/7/28 Laurence Marks L-marks at northwestern.edu The forces are fine. On Jul 28, 2011 4:55 AM, David Tompsett dat36 at cam.ac.uk wrote: Dear All, I have a question about the forces obtained when using TETRA in metallic systems. I read in the 1994 Blochl paper that the Blochl Tetra method is not variational with respect to partial occupancies. Does this mean that the total energy from Tetra is not smoothly variational wrt. to partial occupancies? Therefore, can I trust the calculated forces in a metallic system? Can I use the total energy from Tetra for metals reliably? Many thanks for any help, David Tompsett. ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110728/c8314b8d/attachment.htm
[Wien] Stress
Dear All, Has anyone ever tried to implement the calculation of the stress tensor into Wien? Is there something in the method that prevents its calculation? I read the Nielsen PRB 32, 3780 (1985) where stress calculation is described for local functionals. There are 5 contributions in equation 30 of that paper. Is any in particular difficult for LAPW methods? Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110722/ac4b75b2/attachment.htm
[Wien] Treating d and f states of Eu using SIC
Dear Alay, DFT+U is more than just a SIC correction method. It corrects some self-interaction, but also incorporates Hartree-Fock effects via a Coulomb matrix. See PHYSICAL REVIEW B 79, 035103 (2009). To tread both d and f electrons try: EuTe.inorb 1 1 0nmod, natorb, ipr PRATT,1.0 mixmod, amix 1 2 2 3 iatom nlorb, lorb # treating d electrons with Ueff = 0.5000, How to treat f electrons as well? 1 nsic (LDA+U(SIC) used) 0.500 0.0 U J 0.500 0.0 U J Cheers, David. 2011/7/19 Muhammad Alay-e-Abbas alayabbas at googlemail.com Dear All, I am running wien version 11.1 on a Core 2 duo Laptop with Fedora 8 operating system, ifort compiler and mkl math libraries. The purpose of my calculations is to get band structures of Eu chalcogenides. I want to treat both d and f electrons of Eu atom with the additional U potential using the SIC approach. How shall i modify the case.inorb file if Eu is atom number 1 in the following? EuTe.inorb 1 1 0nmod, natorb, ipr PRATT,1.0 mixmod, amix 1 1 2iatom nlorb, lorb # treating d electrons with Ueff = 0.5000, How to treat f electrons as well? 1 nsic (LDA+U(SIC) used) 0.500 0.0 U J ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110719/fd565a05/attachment.htm
[Wien] how to find Stoner parameter using wien2k
Hi Shamik, Sayed's advice is good. There are also some procedures based on Fixed spin moment calculations. You will have to think carefully about whether your system is metallic or locaised type. The links in the appendix to this paper should help: Ylvisaker and Pickett, PHYSICAL REVIEW B 79, 035103 2009 Best, David. 2011/7/6 Seyed Javad Hashemifar hashemifar at cc.iut.ac.ir I have no experience on calculation of stoner parameter and interesting to learn it. However, by doing a simple spin polarized calculations with nonzero and parallel initial moments on your system, you may easily find whether your system prefers ferromagnetism or not. The initial magnetic moments are controlled in case.inst file. SJ Hashemifar == Seyed Javad Hashemifar Physics Department, Isfahan University of Technology 84156-83111 Isfahan, Iran Tel: +98 311 391 2375 Fax:+98 311 3912376 Email: hashemifar at cc.iut.ac.ir Homepage: http://hashemifar.iut.ac.ir --- 2011/7/6 Shamik Chakrabarti shamikiitkgp at gmail.com Dear wien2k users, According to the Stoner band theory of magnetism, one can able to predict the magnetic ground state of the material as it is paramagnetic or ferromagnetic using product of the non magnetic density of states around the Fermi level (N(E)) and the Stoner parameter (I). As given in some of the literature, the Stoner parameter can be described as the exchange integral and it could be found using LSDA and LMTO methods. see the fallowing literature... 1. http://prb.aps.org/pdf/PRB/v16/i1/p255_1 2. http://jap.aip.org/resource/1/japiau/v89/i11/p6889_s1 We are doing our calculations in Wein2K package using the GGA approximation and we want to know how the stoner parameter could be found according to the my calculations. As non magnetic calculations gives the total density of states at around the Fermi level, it is very helpful for us to predict the material is magnetic or not using the Stoner parameter. therefore, please suggest us how to find the Stoner parameter using wein2k package. -- Shamik Chakrabarti Research Scholar Dept. of Physics Meteorology Material Processing Solid State Ionics Lab IIT Kharagpur Kharagpur 721302 INDIA ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110707/85a8296c/attachment.htm
[Wien] K-point parallel with MPI
Dear All, I have a general question about installing Wien2k 11 without being able to use remote passwordless ssh. In the installation that I have my sysadmin chose yes to the shared memory question in siteconfig. I then have attempted running mixed k-parallel/MPI jobs. What I have found happens is that the mpirun calls for each group of k-points are issued sequentially rather than in parallel, even though I have my .machines file set up correctly. Can anyone advise me as to how I can alter my installation setup so that the MPI calls for each k-point group will be called at the same time? Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110609/18b7efad/attachment.htm
[Wien] Warning in lapw1
Dear All, I have recently compiled the latest Wien2k 11 release, but have come across an unusual warning in case.output1 when running some larger cases (30 atoms or more): :WARN : WARNING: Not all eigenvectors are orthogonal Eigenvalue clusters in eigensolver = 7032153 0 0 0 \ 0 0 0 0 0 0 0 0 \ 0 0 0 0 0 0 0 0 \ 0 0 0 0 0 0 0 0 \ 0 0 0 0 0 0 0 0 \ 0 0 0 0 0 0 0 0 \ 0 0 0 Gap in eigensolver = 1.05489827340396041E-004 -1. -1.\ -1. -1. -1. -1.00\ 00 -1. -1. -1. -1.\ -1. -1. -1. -1.00\ 00 -1. -1. -1. \ -1. -1. -1. -1.\ -1. -1. Increase CLUSTERSIZE in SECLR4 Seclr4(Cholesky complete (CPU)) : 17.200112001.02 Mflops Seclr4(Cholesky complete (WALL)) : 17.251111672.48 Mflops Seclr4(Transform to eig.problem (CPU)) : 84.490 68401.62 Mflops Seclr4(Transform to eig.problem (WALL)) : 85.192 67837.83 Mflops Seclr4(Compute eigenvalues (CPU)) : 213.210 36141.22 Mflops Seclr4(Compute eigenvalues (WALL)) : 213.585 36077.74 Mflops Seclr4(Backtransform (CPU)) : 9.500 28964.20 Mflops This warning appeared in the first iteration of the SCF cycle. I note that the CLUSTERSIZE is set to 600 in SECLR4, but don't understand why I see the warning here. Does anyone know what the cause is? I am running using MPI parallel over 24 cores. Will the parallelisation setup affect this cluster size problem? Thank you, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110603/48350205/attachment.htm
[Wien] Interesting effect of RLO's with Spin Orbit
Dear All, I have been performing calculations with GGA+U with spin-orbit on an Fe and oxygen based material. I have noticed an unusual effect when using RLO's ( the relativistic local orbitals) for the oxygen p-states. I applied the +U orbital potential only to the Fe ion and then applied spin-orbit coupling inside the Fe and oxide spheres. Without using RLO's for the oxygen p-orbitals the DOS shows a good band gap (~3eV). However, when I add RLO's for the oxygen p-states I find that there are states inside the band gap (significant ~0.1 states/eV/FU). The states have the d-character of the Fe ions. It appears that the flexibility introduced by the RLO's on the oxygen are changing the description of the potential such that the groud state has d-orbital character inside the gap. Does anyone have experience with such an effect? I am unsure as to whether the effect is real. The oxygen atoms are light, so I expected the RLO's would have little effect. I did sensible things like converge with respect to the emax in case.in1, k-points, IFFT factor=4. Many thanks for your help, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110421/57ea42a2/attachment.htm
[Wien] MPI Strategy
Dear All, I have recently moved my Wien2k calculations to a new machine (Hector in the UK). This machine does not allow passwordless ssh and therefore disables the k-point parallel part of Wien2k. I therefore want to clarify some points about how MPI (chiefly lapw1_mpi) works in Wien2k. If I run a calculation with say 5 k-points using 5 nodes that each have 24 cores, then will the MPI call of lapw1_mpi split the 5 k-points over the 5 nodes? Or will it simply run the k-points one at a time, using all 120 cores for each one? Is there any documentation I can read about the MPI parallel strategy more generally? Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110401/727edd5b/attachment.htm
[Wien] Runtime problem with lapw1_mpi
Dear All, I have recently been attempting to compile the MPI parallel version of the latest Wien2k. I have successfully generated binaries using ifort (MPI) comiler, linking with MKL 10.2 and FFTW 2. My system uses QLogic MPI. I started some small tests. My lapw0 runs well in parallel, but when lapw1 starts I receive the following error: MPIRUN.node-h18: node-h18:0 node program unexpectedly quit: Exiting. sh: line 39: 5927 Segmentation fault $F_PROG lapw1_1.def Can someone tell me what this $F_PROG variable is? Any hints on what may be causing the issue are welcome. My .machines file looks like: lapw0: node-h18.storage.cluster node-h18.storage.cluster node-h18.storage.cluster node-h18.storage.cluster node-h17.storage.cluster node-h17.storage.cluster node-h17.storage.cluster node-h17.storage.cluster 1:node-h18.storage.cluster node-h18.storage.cluster 1:node-h18.storage.cluster node-h18.storage.cluster 1:node-h17.storage.cluster node-h17.storage.cluster 1:node-h17.storage.cluster node-h17.storage.cluster granularity:1 extrafine:1 lapw2_vector_split:1 Many thanks, David Tompsett. -- next part -- An HTML attachment was scrubbed... URL: http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20110314/6a3b25af/attachment.htm
[Wien] Linearization of Low-lying s-state
Dear All, I am performing calculations on slabs of the material LiFePO4 using GGA+U in the latest Wien2k release. To perform min_lapw internal parameter optimisation I reduce the RMT's and use R=1.06 for Phosphorus. This results in a need to reduce the default for energy separation of core states to -9.2Ry during lstart. As a result Phosphorus 2p states become part of the valence and need to be treated by a LO. I am having trouble with large QTL-B values regarding the L=0 states though. The energy parameters for phosphorus from case.outputst are: P RHFS OCCUPANCYENERGY(RYD) (R4) (R2) (R) (R-1) (R-3) 1S 1.000-1.5314154E+02 5.4020520E-04 1.4448157E-02 1.0359168E-01 1.4648276E+01 1S 1.000-1.5311976E+02 5.3994095E-04 1.579E-02 1.0357948E-01 1.4649640E+01 2S 1.000-1.2760822E+01 1.9082408E-01 3.2223815E-01 5.1970457E-01 2.8247951E+00 2S 1.000-1.2738260E+01 1.8994048E-01 3.2198220E-01 5.1963797E-01 2.8236582E+00 2P* 1.000-9.1715513E+00 1.9086915E-01 2.9430623E-01 4.8389989E-01 2.7216785E+00 2P* 1.000-9.1408359E+00 1.8923948E-01 2.9388805E-01 4.8380643E-01 2.7206626E+00 2P 2.000-9.1041258E+00 1.9405529E-01 2.9690219E-01 4.8614968E-01 2.7052889E+00 6.2902446E+01 2P 2.000-9.0735022E+00 1.9234866E-01 2.9645480E-01 4.8603999E-01 2.7043227E+00 6.2822412E+01 3S 1.000-1.0878541E+00 2.9229783E+01 4.1553776E+00 1.8919617E+00 7.2068651E-01 3S 1.000-8.8775170E-01 3.1604015E+01 4.2438626E+00 1.9036101E+00 7.2402079E-01 3P* 1.000-4.6442261E-01 7.9415784E+01 6.4475288E+00 2.3215625E+00 5.8248088E-01 3P* 0.000-2.7621315E-01 1.1443191E+02 7.4136855E+00 2.4614687E+00 5.6064252E-01 3P 2.000-4.6042138E-01 8.0649286E+01 6.4959005E+00 2.3301759E+00 5.8005227E-01 3.7773200E+00 3P 0.000-2.7242986E-01 1.1708022E+02 7.4892061E+00 2.4735418E+00 5.5768379E-01 3.5655831E+00 Examples of the low lying states I find in case.output1up and case.output1dn are:
[Wien] problem with DOS calculations
Dear Fhokrul, With only one k-point you should only have a discrete set of energies. Therefore I think the DOS would be a set of delta functions. Your energy resolution in case.int is unlikely to capture them. Best, David. Md. Fhokrul Islam wrote: Dear Wien2k users, I am trying to calculate DOS for a surface supercell but I am not sure why case.dos1evup file shows only zeros in the columns for density of states. I have calculated DOS many times before without any problem for bulk system. In my surface calculation I have used MPI version of wien2k and have only one k-point. I have used the following steps as usual, x lapw2 -c -qtl -p -up x lapw2 -c -qtl -p -dn edited case.int file x tetra -up x tetra -dn I have tried different range of energies in case.int file but couldn't make it work. I would appreciate if anyone can tell me how to solve this problem or whether there is a problem in calculating DOS for only one k-point. Thanks, Fhokrul Hotmail: Trusted email with powerful SPAM protection. Sign up now. https://signup.live.com/signup.aspx?id=60969 ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] Fermi surface in XCrySDen with data from wien2k parallel calculations
Dear Jian-Xin, The Fermi surface can not be plotted from a parallel calculation because XCrysden needs the eigenvalues from a single case.output1* file. From a parallel calculation a separate case.output1* file is produced for each cpu. So to solve the problem you need to write a small script (eg. using awk) that combines the separate case.output1* files into one file. Let me know if you have problems with it. Best, David. Jian-Xin Zhu wrote: Dear Prof. Blaha and Wien2k users, I am trying to plot the Fermi surface in XCrySDen for a system with spin-orbit coupling. After I finish the self-consistency calculations by run_lapw -so -cc 0.0001 -i 40 -p I save the data and start the following to prepare the data for Fermi surface plotting with XCrySDen (k points have been prepared without shift as required by XCrySDen) --- x lapw1 -p x lapwso -p x lapw2 -so -fermi -p I then start the XCrySDEN and select File --- Open WIEN2k --- Fermi Surface (hit OK) --- Render Fermi Surface I then got error in reading the data. Note the panel of XCrySDen: Fermi Surface Preparation does not provide the case to run the calculations with spin-orbit coupling, that is, x lapw2 -so -fermi If I take off the -p option above, and I can plot the Fermi surface successfully. It makes me wonder whether I have missed some steps when -p option is used. I do notice that the file case.energy is empty when x lapw1 -p is executed. So I try running x sumpara -d after finishing x lapw2 -so -fermi -p, to no avail. I searched through the Wien2k archive and could not find a discussion on this issue. Your help/advice is appreciated. Jianxin ___ Wien mailing list Wien at zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] Missing lapwdmc
Dear All, After recently updating to the latest version of Wien2k I found that the lapwdmc executable had not been produced, but the real version had been. I copy the compile.msg below from SRC_lapwdm/. Does anyone know how I can fix this problem? Thank you, David. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] Problem Energy Sep LAPW2 ACTUALLY MKL PROBLEM
Dear All, I recently wrote to the list facing problems with strange eigenvalues leading to errors in lapw2. I have solved the issue and thought I would write to the list as the actual cause is due to errors in MKL 10.1 and early 10.2 documented at: http://software.intel.com/en-us/articles/dgemm-and-sgemm-accuracy/ I am not sure if these problems have been mentioned on the list before, but may affect anyone who has built Wien2k with any of the affected versions of Intel MKL. The errors only occur when the system size is such that the matrix sizes in lapw1 satisfy the criteria listed at the link above. Thus it may appear that one's code is working well and then odd errors will occur for particular cases. After rebuilding my Wien2k with either old MKL 9.1 or the latest 10.2.2.025 the problems disappeared. Best wishes, David Tompsett. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/ Peter Blaha wrote: You said, this happens already in the first iteration ? What are your RMT values of all atoms ? RKmax ? Are you sure, your structure is correct ? The message 15 eigenvalues below -9. indicates that much more is incorrect than just the energy parameters of atom 24. I don't think you should put the E-parameters to -8 Ry. Probably you do not expect eigenvalues in this range. Hard to guess what you could have made wrong. David Tompsett schrieb: Dear Peter and All, Thank you for the help. From the case.output2 files it seems that the problem is with very low lying p-states: QTL-B VALUE .EQ. 1876.88467 in Band of energy -8.53431 ATOM= 24 L= 1 Check for ghostbands or EIGENVALUES BELOW XX messages Adjust your Energy-parameters or use -in1new switch, check RMTs !!! My linearization energies for this atom are: 0.605 0 (GLOBAL E-PARAMETER WITH n OTHER CHOICES, global APW/LAPW) 00.60 0.000 CONT 1 0 -4.00 0.005 STOP 1 1 -2.21 0.010 CONT 1 10.60 0.000 CONT 1 20.60 0.010 CONT 1 My first thought is to alter -2.21 to -8.00 for the p-state LO, but I am concerned that since -8.53431 is so low that somehow my calculation window is including multiple p-states ie. more than two sets of p-states. Also, the case.output1 file says that there are: 15 EIGENVALUES BELOW THE ENERGY -9.0 Please advise me as to how to proceed. Thank you, David. Peter Blaha wrote: You are doing k-parallel calc ! What says lse ? Check the individual case.output2_xx files ! David Tompsett schrieb: Dear Peter and All, the end of my case.output2 looks as follows and confirms that there is some error in finding the separation of semi-core and valence states, though from the userguide it seems that this should not be a major issue without using -in1new. eseper below EF 0.500eseper minimum gap 5.000E-002 Energy to separate semicore and valence states: -999. :NOE : NUMBER OF ELECTRONS =1313.000 :FER : F E R M I - ENERGY(TETRAH.M.)= 0.94199 NbFe2Super222NM.weigh 8 writing NbFe2Super222NM.weigh_1 writing NbFe2Super222NM.weigh_2 writing NbFe2Super222NM.weigh_3 writing NbFe2Super222NM.weigh_4 writing NbFe2Super222NM.weigh_5 writing NbFe2Super222NM.weigh_6 writing NbFe2Super222NM.weigh_7 writing NbFe2Super222NM.weigh_8
[Wien] Problem with Energy Separation in LAPW2
10351.96425840761.9977970344 0.00D+00 band10361.96677832661.9948459103 0.00D+00 band10371.97010746941.9964735701 0.00D+00 band10381.97142902331.9942832855 0.00D+00 band10391.97191115251.9950427054 0.00D+00 band10401.97382520671.9990564146 0.00D+00 band10411.97673260211.9980496124 0.00D+00 band10421.98318581341.9831858134 0.00D+00 band10431.98793378881.9879337888 0.00D+00 band10441.99082813551.9908281355 0.00D+00 band10451.99389282781.9938928278 0.00D+00 eseper below EF 0.500eseper minimum gap 5.000E-002 Energy to separate semicore and valence states: -999. :NOE : NUMBER OF ELECTRONS =1313.000 :FER : F E R M I - ENERGY(TETRAH.M.)= 0.94199 NbFe2Super222NM.weigh 8 writing NbFe2Super222NM.weigh_1 writing NbFe2Super222NM.weigh_2 writing NbFe2Super222NM.weigh_3 writing NbFe2Super222NM.weigh_4 writing NbFe2Super222NM.weigh_5 writing NbFe2Super222NM.weigh_6 writing NbFe2Super222NM.weigh_7 writing NbFe2Super222NM.weigh_8 NK 8 Peter Blaha wrote: Please check case.output2 D.A. Tompsett schrieb: Dear All, I am trying to run a supercell calculation of transition metals in a Laves AB2 type phase. The supercell contains 96 atoms and one atom of type B doped onto the site of an A atom. My calculation is repeatedly failing in LAPW2 and the stderr shows a QTL-B error. My case.scf2 looks like: Can anyone suggest to me how I might fix this problem or where I should look to solve it? Many thanks, David Tompsett. Peter Blaha wrote: Please check case.output2 D.A. Tompsett schrieb: Dear All, I am trying to run a supercell calculation of transition metals in a Laves AB2 type phase. The supercell contains 96 atoms and one atom of type B doped onto the site of an A atom. My calculation is repeatedly failing in LAPW2 and the stderr shows a QTL-B error. My case.scf2 looks like: Can anyone suggest to me how I might fix this problem or where I should look to solve it? Many thanks, David Tompsett. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] Problem with Energy Separation in LAPW2
Dear Peter and All, Thank you for the help. From the case.output2 files it seems that the problem is with very low lying p-states: QTL-B VALUE .EQ. 1876.88467 in Band of energy -8.53431 ATOM= 24 L= 1 Check for ghostbands or EIGENVALUES BELOW XX messages Adjust your Energy-parameters or use -in1new switch, check RMTs !!! My linearization energies for this atom are: 0.605 0 (GLOBAL E-PARAMETER WITH n OTHER CHOICES, global APW/LAPW) 00.60 0.000 CONT 1 0 -4.00 0.005 STOP 1 1 -2.21 0.010 CONT 1 10.60 0.000 CONT 1 20.60 0.010 CONT 1 My first thought is to alter -2.21 to -8.00 for the p-state LO, but I am concerned that since -8.53431 is so low that somehow my calculation window is including multiple p-states ie. more than two sets of p-states. Also, the case.output1 file says that there are: 15 EIGENVALUES BELOW THE ENERGY -9.0 Please advise me as to how to proceed. Thank you, David. Peter Blaha wrote: You are doing k-parallel calc ! What says lse ? Check the individual case.output2_xx files ! David Tompsett schrieb: Dear Peter and All, the end of my case.output2 looks as follows and confirms that there is some error in finding the separation of semi-core and valence states, though from the userguide it seems that this should not be a major issue without using -in1new. eseper below EF 0.500eseper minimum gap 5.000E-002 Energy to separate semicore and valence states: -999. :NOE : NUMBER OF ELECTRONS =1313.000 :FER : F E R M I - ENERGY(TETRAH.M.)= 0.94199 NbFe2Super222NM.weigh 8 writing NbFe2Super222NM.weigh_1 writing NbFe2Super222NM.weigh_2 writing NbFe2Super222NM.weigh_3 writing NbFe2Super222NM.weigh_4 writing NbFe2Super222NM.weigh_5 writing NbFe2Super222NM.weigh_6 writing NbFe2Super222NM.weigh_7 writing NbFe2Super222NM.weigh_8 NK 8 Otherwise the file seems OK and I copy the rest of it below (excluding the KVEC list). The error is occurring on the very first SCF cycle each time. Many thanks for any more help, David. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] lcore Stops
Dear All, I am having a problem while attempting an scf calculation of Ag5Pb2O6. In setting up the calculation the default RMT's mean that there is core charge leakage from the large Pb atom if one uses the default separation energy of 6.0Ry. I have tried to get around this in two ways: 1) Set the separation to 8.0Ry. The lstart indicates no leakage. 2) Increase the RMT of Pb (and decrease that of O to compensate) so that even with 6.0Ry separation input to lstart there is no charge leakage. In both cases lcore fails in the same way. The program lcore fails right after it is called leaving no output or error file that I can find. lapw2 seems to have produced all of its output files successfully. In case (2) above I also attempted to reduce RKmax to adjust for the smaller value of RMTmin, but the same type of error occurs. Is there something else that I should be doing in the setup? Or does anyone know where I should look to find the relevant error? Thank you, David. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/
[Wien] Checking L,M Basis for new Local Rotation Matrix
Dear All, I have changed the local rotation matrix for an atom in my system from the default found by program SYMMETRY. I have determined what I think the new L,M basis should be according to the prescription in reference Kara and Kurki-Suonio 81. In the UG it says Perform a run of LAPW1 and LAPW2 using the GAMMA-point only (or a complete star of another k point). In such a case, ?wrong? LM combinations must vanish. I want to check my new basis by this method. In what sense would the wrong LM vanish? Which output file would I look at? Thank you for any help, David. -- David A. Tompsett Quantum Matter Group Cavendish Laboratory J. J. Thomson Avenue Cambridge CB3 0HE U.K. Tel: +44 7907 566351 (mobile) Fax: +44 1223 768140 http://www-qm.phy.cam.ac.uk/