[Pw_forum] Mean internal potential for e-microscopy
Dear users and developers, I have two related questions about what would be more meaningful to extract from a SCF calculation with DFT in Quantum Espresso as far as the *mean internal potential* is concerned. This quantity, in electron microscopy experiments, is linearly related (at least in first approx) to the response of the electron beam passing through a sample. In developing the theory of the response one uses the single particle static electic potential felt by the electrons given, in turn, by the solution of the Poisson equation for the total charge existing in the interior of the sample. Now: a. I wonder if the the DFT setting one has to use the Kohn-Sham potential (as "best" single body potential in the many-body problem) or only the contributions coming from ions + Hartree term coming from the electronic charge. In a series of papers of the '90s o 2000s I have seen that they use the second option. The rationale would be that the electrons from the beam at high energy are not really a part of the sample and are not correlated with the ones of the interior. Would you agree? b. Again in those papers they choose *linear augumented plane wave full potential*. Why should one use PAW? In which sense full potentials? Aren't Pseudos sufficient? Thanks a lot for your time. -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/ ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] Error in routine read_rho_xml
Dear all, in am writing here for suggestions about a (post) post-processing issue with the executable pp.x I have some (supposedly complete) outputs from relaxations carried out with pw.x in a large parallel machines, now stored as compressed files. If I try to decompress the output directory (in which I find *.wfc* files and output directory with prefix.save subdirectory) then run pp.x (with some specific postprocessing input file) I get the error "Error in routine read_rho_xml (1) cannot open ./out/acm15-df2.save/charge-density.dat file for reading" because typically the charge-density.dat file or even the file vdW_kernel_table cannot be read, despite the fact that I can correctly list them using the path specified in the error lines and that I set all the r/w permissions on the decompressed directory. Any clues? Thanks for your time, Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/ ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] The unit of potential (??)
Hi Evan, regarding your question below, various times I had to extract the potential with the same parameters for pp.x AFAIK the units used in that case are ATOMIC RYDBERG UNITS, that means Rydbergs (Ry) for energy, and 1 Ry = 13.6057 eV Cheers >Hello, everyone >I output the potential by pp.x (iflag = 3, output_format = 5), I like >to know what is the unit of the potential, and how to convert to the >unit with eV. >Thank you in advance. >Evan >USC, China -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/ ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] Fwd: PhD opportunities at SIMaP, Grenoble
Sorry for possibile double postings about the positions below. --- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/unit/users/degliesposti -- Forwarded message -- From: ROBERTA POLONI Date: Wed, Apr 12, 2017 at 12:02 PM Subject: PhD opportunities at SIMaP, Grenoble To: Roberta Poloni Dear colleagues and friends, if you know of motivated students that might be interested please share these two PhD openings: Thanks a lof Roberta topic 1: Computational study of MOFs for an efficient carbon capture Applications are invited for a fully funded 3-years PhD position to work on the electronic structure modeling of porous coordination polymers starting on OCTOBER 1ST 2017 and funded by the ANR ComputationalCarbonCapture. We propose to design novel families of MOFs whose high affinity for CO2 can be modified under light treatment or high temperature, so that adsorption and desorption are performed each at the most convenient conditions. The idea is that the high heat of adsorption will allow for a large uptake, while the lower affinity for CO2 upon light or temperature treatment will allow for a lower regeneration energy and therefore a lower energy penalty. The work will involve the understanding of the mechanism by which experimentally synthesized photoactive materials composed by azo-benzene functionalized metal-organic frameworks are capable to selectively adsorb certain gases, in particular CO2 over N2. This will involve the study of the thermodynamics of gas uptake as well as the study of the isomerization mechanism of the light sensitive molecule attached to the MOF. Due to the poor performance of these alrady-synthesized materials, we seek to predict how these should be modified in order to enhance the efficiency of the process. We will also study spin crossover MOFs with scope of computationally designing specific MOFs whose affinity for CO2 (choise of metallic nodes and ligands) and spin crossover transition (temperature, hysteresis) make them good candidate for efficient carbon capture materials. The PhD candidate will learn to use several methods for this project, i.e. DFT, TDDFT, GW/BSE and classical molecular dynamics and Monte Carlo methods. The thesis will be performed at the SIMAP laboratory (http://simap.grenoble-inp.fr/) in Grenoble (France) under the supervision of ROBERTA POLONI, CNRS, (http://www.robertapoloni.com/) but the PhD candidate is expected to interact also with other researchers involved, CLAUDIO ATTACCALITE, CNRS, (http://www.attaccalite.com/ Marseille, France) and LI-CHIANG LIN (OSU, https://cbe.osu.edu/people/lin.2645 [1]). The applicants should have a University degree in physics, chemistry, materials science, engineering or similar. To apply please send an application cover letter explaining briefly your motivation and interest in this position to roberta.pol...@simap.grenoble-inp.fr, also a CV including the transcript of your undergraduate degree, and name and e-mail address of two academic referees. - topic 2: COMPUTATIONAL DESIGN OF ULTRA-LOW K MOFS Applications are invited for a fully funded 3-years PhD position to work on the modelling of dielectric properties of MOFs, starting on OCTOBER 1ST 2017. The project aims at computationally designing advanced dielectrics for microelectronics. For several decades, silica-based insulators have been developed and optimised to both insulate the metal interconnections and avoid parasitic effects like crosstalk induced by capacitive coupling. The key magnitude for the latter phenomenon is the dielectric constant (k-value) that has been reduced down to 2.3-2.5 by tuning the chemistry and porosity of these insulators. Because of the continuous device miniaturisation, the trade-off between k-value, stiffness, leakage, integration process,… gets more and more difficult to obtain with these historical materials. We propose to develop a new strategy for the design of novel ultra-low-k candidate materials by combining computational and experimental studies. Specifically, we propose to study Metal-Organic Frameworks (MOFs) since their extraordinary chemical tunability and porosity will allow to explore a myriad of materials whose density, topology and chemistry can be fine-tuned in order to design a material with a specific dielectric constant in the ultra-low range (2.0http://www.robertapoloni.com/) at the SIMAP laboratory (http://simap.grenoble-inp.fr/) in GRENOBLE (France) and in collaboration with JIHAN KIM at KAIST (http://cbe.kaist.ac.kr/ab-1110-5 [2]). He/She is expected to strongly interact with the experimental group led by FABIEN VOLPI at SIMaP, INPG (fabien.vo...@simap.grenoble-inp.fr). The applicants should have a University degree in physics, chemistry, materials science, engineering or similar
[Pw_forum] Errors in last step of GW calculations with GWL
Dear all, I am writing you because (together with a colleague of mine Andrea Pedrielli) we are experiencing errors in the last step (gww.x) of a GW calculation on graphene with the code GWL. QE version is 6.1 In particular, with different choices of parameters for the polarization basis and frequency grids, we get either FCN: MAXN TOO SMALL or Routine fit_multipole: chi1 > chi0 but we haven't found clues in the documentation to get rid of these errors. Any help? Thanks for your time -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/ ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] Display of forces in a pure scf run
Dear developers, I have a very basic question, that I had never realized so far and has been asked by a students of mine... If I set to .TRUE. the logical fields tstress and tprnfor in the &CONTROL stanza of the pw.x input during a pure scf calculation, how are the forces on the ions calculated? What do the displayed values represent if only a scf calculation is performed and the ionic coordinates are not allowed to change? Thanks for your time. -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 degliesposti -AT- bo.imm.cnr.it
[Pw_forum] Display of forces in a pure scf run
Dear Stefano and Axel, thanks for your replies. Actually question can be rephrased as follows. Thanks to the Hellmann-Feynman theorem in order to compute the ionic forces I need only the partial derivatives of the total energy with respect to the ionic coordinates (see for instance 7.5 and 7.8 in Payne et al., RMP 64, p. 1045). If no real displacements of the atoms are needed to compute these derivatives, then I guess that they can be simply calculated as expectation values of the Hamiltonian on the current wavefunction because they can written down once a (parametrized) expression of the ionic potential is chosen. Am I right? Thanks for still devoting time to this... -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 degliesposti -AT- bo.imm.cnr.it
[Pw_forum] Choice of irreducible representations with ph.x
I would like to perform a phonon calculation at Gamma making a choice of only a selected number of irreducible representations. After a proper scf calculation I use with ph.x with an input file like the following one: Phononic spectrum (1) of CoM (at Gamma) &inputph amass(1)=1.00794, amass(2)=12.011, amass(3)=14.007, amass(4)=15.999, amass(5)=55.845, prefix='CoM', outdir='/scratch/userinfm/cne0fm4a/CoM/results/', iverbosity=1, epsil=.false., fildyn='CoM.dm', nrapp=1, / 0.0 0.0 0.0 1 The problem is that the calculation seems to proceed ignoring the nrapp value and the list of numbers I put in the last line and all the irreps are considered. This happens both for nrapp=1 (one value in the list) or nrapp>1 (nrapp values in the list). pw.x and ph.x are run on the CINECA cluster where there appears to be QE ver. 4.1CVS and I would liketo save cpu time by excluding the undesired irreps. Using the same syntax (nrapp and the list of irreps at the end) to modify the scripts of example02 I successfully select only one of the two irreps for Si with QE ver. 3.2 installed on a desktop pc (reporting the message "Not done in this run"). Might it be a version issue? Thanks a lot for your help and time. -- ___________ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Compilation of QE 4.3.2 on (grid) cluster with OpenMPI / OpenMP
Dear all, together with Roberto Alfieri in Parma we are trying to install QE 4.3.2 on a new cluster to be used also in a Grid environment. We are experiencing two types of problems: 1) if one selects compilation with OpenMP (--enable-openmp) the following error is reported: # sum_band.f90:121: internal compiler error: in make_decl_rtl, at # varasm.c:893 2) if OpenMP is not selected compilation seems to be ok, but then the following errors are reported when one tries to run examples involving executables other than pw.x running the band-structure calculation for Si... ### # FROM IOTK LIBRARY, VERSION 1.2.0 # UNRECOVERABLE ERROR (ierr=1) # ERROR IN: iotk_scan_end (iotk_scan.f90:241) # CVS Revision: 1.23 # foundl # ERROR IN: iotk_close_read (iotk_files.f90:746) # CVS Revision: 1.20 ### running the phonon calculation at X for Si... ### # FROM IOTK LIBRARY, VERSION 1.2.0 # UNRECOVERABLE ERROR (ierr=1) # ERROR IN: iotk_scan_end (iotk_scan.f90:241) # CVS Revision: 1.23 # foundl # ERROR IN: iotk_close_read (iotk_files.f90:746) # CVS Revision: 1.20 ### The Linux environment essentially is Scientific Linux 5.4 x86_64 with gcc 4.1.2 and openmpi-1.4-4 Does anyone have suggestions on how to fix these errors? Thanks for your time. Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 degliesposti -AT- bo.imm.cnr.it
[Pw_forum] Problems with XSpectra code
Dear all, I've installed QE 4.1.2 to try the XSpectra code. The input file is as follows &input_xspectra calculation='xanes_dipole', prefix='MbCO_scf', outdir='/home/desposti/temp/MbCO/out/', verbosity='high', xiabs=1, / &plot / &pseudos filecore='Fe.wfc' / &cut_occ / 1 1 1 0 0 0 After some minutes xspectra.x stops (EXIT 2) with the following message %% from main program : error #16 Wrong xiabs!!! %% Before the message it seems that the file of the previous scf run is correctly read (at least the crystal structure). This run is made with a pseudpotential Fe.pbe-sp-van_ak.UPF for the absorbing Fe atom placed in first position in the input file (so it should be xiabs=1 right?) I wonder if the error is related to the fact that when I generate the core wavefunction Fe.wfc using ./upf2plotcore.sh Fe.pbe-sp-van_ak.UPF > Fe.wfc in the file Fe.wfc I simply get # number of core states 0 = The same happens to me if I try the Fe pseudo in the QE repository called Fe.star1s-pbe-sp-mt_gipaw.UPF. For the moment I don't want core-hole effects; it's just to try... So, what could it be the cause of the "xiabs" error above? Thanks for your time. Cristian ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] little 'bug' in bands_FS
Dear developers, I have just noticed a little 'bug' in the code bands_FS for Fermi surfaces visualisation, in the sense that it was probably written and tested to read correctly the usual (text) file coming out from previous bands/nscf calculations with pw.x. But if the bands/nscf calculation is performed with the option of *high verbosity*, then in the output file there appear occupation numbers just after the list of Kohn-Sham energies for a given wavevector and the number of desired bands. As a result in this case (i.e. high verbosity) the file is not read correctly from bands_FS.x and one gets a read error. I guess that the solution would be either (i) to rembember to exclude verbose output in the preliminar bands/nscf calculation or (ii) add a little control within bands_FS in order to skip the occupation numbers lines while reading the energies from the file. Thanks for your time -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] Question on last line of XSpectra input file
Dear all, if I have not missed something, in the XSpectra input documentation as well as in the examples I do not understand the precise meaning of the last line after closing the &cut_occ stanza. It seems like a repetition of the K_POINTS specification in the input for pw. For instance if the latter is set to be "automatic" with "1 1 1 0 0 0" I guess it means a 1x1x1 grid in k-space with zero offset, and then I repeat the same in the XSpectra input file. But what should I specify in the latter file if in the scf calculation with pw I use other options for K_POINTS, for instance Gamma or a list of points with weights? I hope the question is clear. Thanks for your time. Cristian -- _______ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] greetintgs
Thanks for the really nice message. Simply I point out that most of the Christians (like myself) would agree that the most important Christian festivity is Easter, the resurrection day. So, having passed Christmas, let me wish you all a great 2010! Cristian > Message: 4 > Date: Thu, 24 Dec 2009 16:45:19 +0100 > From: Stefano Baroni > Subject: [Pw_forum] greetings > To: PWSCF Forum > Message-ID: > Content-Type: text/plain; charset="iso-8859-1" > > Not sure what would be worse: not to greet anybody, or send non-Christians > (like myself) greetings for the most important Christian festivity? Don't > know about you, but I love to be greeted on Christmas even if I am no > Christian. So, please all accept my humble greetings for a > > Merry Christmas and a Happy New Year > > Stefano > > PS - the bottom line is political correctness is a rather pedantic exercise! > > --- > Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste > http://stefano.baroni.me [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni > (skype) > > La morale est une logique de l'action comme la logique est une morale de la > pens?e - Jean Piaget > > Please, if possible, don't send me MS Word or PowerPoint attachments > Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -- ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Choice of irreducible representations with ph.x
Here follows the input file for the preliminary scf calculation done using pw.x: Thanks, Cristian --- &CONTROL calculation = "scf", prefix = "CoM", pseudo_dir = "/scratch/userinfm/cne0fm4a/pseudo/", outdir = "/scratch/userinfm/cne0fm4a/CoM/results/", verbosity= "high", / &SYSTEM ibrav = 0, celldm(1) = 1.8897261D0, nat = 48, ntyp = 5, ecutwfc = 100.D0, ecutrho = 400.D0, nbnd = 100, / &ELECTRONS electron_maxstep = 500, conv_thr= 1.D-6, mixing_beta = 0.7D0, diagonalization = "david", / CELL_PARAMETERS cubic 20.0 0.0 0.0 0.0 20.0 0.0 0.0 0.0 15.0 ATOMIC_SPECIES H 1.00794 H.pbe-van_ak.UPF C 12.011 C.pbe-van_ak.UPF N 14.007 N.pbe-van_ak.UPF O 15.999 O.pbe-van_ak.UPF Fe 55.845 Fe.pbe-sp-van_ak.UPF ATOMIC_POSITIONS {angstrom} Fe 0.0 0.0 0.0 0 0 0 N -2.015936221 0.0 0.038768204 N -0.13925 -2.014702715 0.046905328 N2.017269908 0.0 0.040458155 N -0.13925 2.014652715 0.046905328 C -2.847040814 1.101001374 0.021528526 C -2.847040814 -1.101001374 0.021528526 C -1.099901325 -2.845127136 0.035557609 C1.099672879 -2.845475866 0.033358015 C2.848427407 -1.101716959 0.018059959 C2.848427407 1.101716959 0.018059959 C1.099672879 2.845475866 0.033358015 C -1.099901325 2.845127136 0.035557609 C -2.424089030 2.424582315 0.023330214 C -2.424089030 -2.424582315 0.023330214 C2.424273228 -2.425359652 0.019740116 C2.424273228 2.425359652 0.019740116 C -4.226164834 0.682755749 0.013419289 C -4.226164834 -0.682755749 0.013419289 C -0.682749799 -4.225087959 0.031048576 C0.682057949 -4.225434105 0.029188709 C4.226987371 -0.683186680 0.002120738 C4.226987371 0.683186680 0.002120738 C0.682057949 4.225434105 0.029188709 C -0.682749799 4.225087959 0.031048576 H -3.194929270 3.196716641 0.005583502 H -3.194929270 -3.196716641 0.005583502 H3.194370362 -3.198147171 -0.001036024 H3.194370362 3.198147171 -0.001036024 H -5.076004919 1.361487055 0.008732438 H -5.076004919 -1.361487055 0.008732438 H -1.361920855 -5.074396365 0.029488844 H1.360598646 -5.075191580 0.024602540 H5.077216872 -1.361288276 -0.011051138 H5.077216872 1.361288276 -0.011051138 H1.360598646 5.075191580 0.024602540 H -1.361920855 5.074396365 0.029488844 N -0.005537284 0.0 2.101372675 C1.071489062 0.0 2.880387333 C -1.100816954 0.0 2.946533090 N0.704089722 0.0 4.188705635 C -0.674259140 0.0 4.251114030 C -0.006438108 0.0 -1.731247148 O -0.017672814 0.0 -2.897017310 H -2.113804696 0.0 2.562350854 H2.100245178 0.0 2.539608359 H1.339496592 0.0 4.980002249 H -1.211591117 0.0 5.192219697 K_POINTS 1 0.0 0.0 0.0 1.0 -- _______ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Choice of pseudpotential (d orbitals for Fe)
Dear users and developers, I'm using Quantum Espresso 3.2 and would like to reproduce some of the DFT results of C. Rovira et al., J. Phys. Chem. A 101, 8914 (1997) regarding the equilibrium geometries of Iron-Porphyrin (FeP) complexes. However, I find some significant discrepancies: in particular in our equilibrium structure of FeP(Im)(CO), where Im denotes the imidazole group, the FeP plane gets considerably more "domed" with respect to what reported in the above-mentioned reference. Of course there could be several reasons for this, but in particular I noted that they use XC functional of Becke (exchange) and Perdew (corr) type. They also mention a comparison with BLYP-type functionals. However, since orbital energy structure of the ground state involves d orbitals of Fe, initially I chose the pseudopotentials (in the pwscf site) that include the d orbitals contribution, namely Fe.pbe-nd-rrkjus.UPF or Fe.pz-nd-rrkjus.UPF For the sake of uniformness I used PBE- or PZ-type pseudopotentials also for C, O and N. Now, I wonder if the problem of excessive doming could be due to the use of wrong pseudopotential ... I would like to try with BLYP's but the one for Fe does not seem to include explicitly d orbitals. Are there other sources of pseudopotentials? Thanks in advance for any help. Cristian ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] XSpectra - tail of the spectrum falling relatively quickly
Dear Yevgen, I would like to contribute my little experience with XSpectra to answer your question about the difference between computed and experimental spectra at high energies. In fact it is a problem we have encountered in all the cases we have studied with a student of mine (Emiliano Burresi), both for test systems and for nontrivial clusters of metalloproteins. I made the same questions about four years ago to Matteo Calandra, the main developer of the XSpectra module, and I translate below the essential part of the answer... "At high energies (in the EXAFS region) the theoretical curve tends to zero while the experimental one doesn't. The reason lies in the way the projectors of the PAW method are built in Quantum Espresso; they have no nodes and as such they are no longer a valid basis at high energies. The decay to zero of the theoretical curve is an artifact, actually it should settle to a constant. For this reason XSpectra allows for a calculation of XANES (X-ray absorption NEAR EDGE structure) but not EXAFS." With Emiliano we simply considered a test case (Silicon) and tried to increase the number of projectors in the PAW scheme; as shown in the figure here (not published) https://dl.dropboxusercontent.com/u/47323550/XANES_Si.tif the spectrum goes at least in the right direction, so this may be a viable solution even if not particularly efficient. Now, another question is if the situation has changed since 2010, and I cannot be more precise on this. Hope this helps. Cristian -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] Quantum Espresso (4.3.2) and pure openmp runs
Dear all, I have a (maybe basic) question on the performance of Quantum Espresso with respect to openmp, even if I do not realise if the answer lies in the way QE handles different levels of parallelism or lies at the system level instead. However, I am running QE 4.3.2 on a quad-core machine with 2 CPUs 64bit. SO is Debian stable, and it recognises 16 threads in the CPU table. I have compiled QE with pure openmp support (gcc) without any MPI environment. The question is: even if I do NOT set the OMP_NUM_THREADS variable at runtime, from the table of the "top" command my process now uses 1000 % CPU time, as if it were "parallelised" over 10 threads. Is this real? Is there some kind of default when pw.x is invoked? If so, how is openmp parallelism exploited by QE? Thanks for your time and support. -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] Quantum Espresso (4.3.2) and pure openmp runs (Axel Kohlmeyer)
Thanks for the reply regarding pure OpenMP runs, it clarifies a lot. -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] Minor problem in tests and cptests scripts for QE 4.3
Dear users, I've just installed Quantum Espresso 4.3 and performed the tests and cptests. However (contrarily to what written in the comments within the scripts), in order to get them work I had to define the environment variables TMP_DIR and PSEUDO_DIR within the scripts check-pw.x.j and check-cp.x.j because they are not defined in each input file. Am I missing something or is it a minor problem with the scripts? Thanks for your time, Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 degliesposti -AT- bo.imm.cnr.it
[Pw_forum] Units in dynamical matrix files
Dear all, I would like to have some clarifications on the units used in dynamical matrix files produced by the ph.x program in the QE suite. At the beginning of dynamical matrix file the masses are not expressed in the same units as in the input files used for pw.x, that is [amu]. It seems to me that in the dynamical matrix file the masses could be in atomic units, which I would say correspond to electron masses. So there should be a conversion factor 1 amu = (1.660538782 ? 10^-27 kg / 9.1093826 ? 10^-31 kg) me = 1822.8884 me However there is still a factor 1/2 to include to pass from the amu values to those reported at the beginning of the dynamical matrix file. Am I right? Why the 1/2 factor? And what are the typical units used for force constants? Thanks a lot. Cristian -- ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Symmetry of the dynamical matrix
Dear all, using the phonon code I noticed that, even for simple molecular structures, the dynamical matrix computed by ph.x is not symmetric. (I limited myself to Gamma point calculations.) In general, this does not guarantee real eigenvalues. How are the eigenfrequencies computed then? Thanks for your time, Cristian -- ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Additional informations about XSpectra
Dear all, we are planning to use XSpectra and therefore I would need to understand in detail the following three elemets of the input: (1) is the energy range (xemin,xemax) referred to the Fermi energy of the preliminar scf calculation? (2) is there a special convention for the components of the wavevector and of the polarization vector? The defaults are (1,0,0) and (0,1,0); shouldn't they be orthogonal? (3) last line of input, after the $cut_occ section (question already posted on past 2/7 in this forum) Thanks for your time. Cristian -- ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Additional informations about XSpectra - Peaks at negative energies?
Thanks to Matteo Calandra for his answers. Regarding the Fermi level, I have now an additional question based on an example we are dealing with. With a preliminar run of XSpectra I see a Fermi energy of -0.259, so I use this value for ef_r in following calculations. If I include negative energies (say -30 eV, with respect to this one) in the input file then in the resulting spectrum I see a *sharp* peak at about -5 eV. Now, if levels at negative energy should be all occupied, how can it result such an evident absorption peak? Is it related to the "cut_occ_states" option? By default it is .false. What does this mean exactly? Thanks again for your help. Cristian > (1) is the energy range (xemin,xemax) referred to the Fermi energy > > of the preliminar scf calculation? > > No, it is referred to the Fermi level given in input in the xspectra > file (input variable ef_r in the input_xspectra namelist. This means > that if you do not specify the Fermi level then ef_r=0.0 and > they are referred to the energy zero of the preliminary scf run. -- ___________ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Installing QE 4.0.5 on Mac Pro 64-bit and gfortran
>Have you tried with 4.3.x for instance (not Apple build)? There should >be also 4.4.x available. I have just tried to compile with gcc 4.3.3 (GCC build). The configure step is like this: sudo ./configure --disable-parallel CC=gcc-4 CFLAGS="-m64" FFLAGS="-m64" F90FLAGS="-m64" CPPFLAGS="-m64" LDFLAGS="-m64" (gcc-4 means 4.3.3 on our Mac Pro). When I try 64-bit compilation the errors with fft_scalar are still there... (for instance "Symbol ... at (1) already has basic type of INTEGER"). I searched in Google for this type of errors with gfortran and it is likely that Axel Kohlmeyer is right: >ok. now that explains something. your default compiler installation >is a 32-bit version. those have a limited address space of about >2GB. this is why you are running out of memory. when you change >to 64-bit mode, this is resolved. however, you cannot just change >make.sys. you have to re-run configure and pass the -m64 flag >with the compiler flags (for c, f77 _and_ f90) and then compilation >should work. in 64-bit mode some parts of the fortran and c code >in Q-E have to be adjusted due to the difference in pointer size >vs. the default integer size compared to 32-bit mode. So, where and how should I fix the pointer size vs. the default integer size? Thanks again. Cristian -- ___ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] QE 4.0.5 64-bit serial installed on Mac Pro with gcc 4.3.3 and gfortran
Dear all, I only want to report briefly that eventually it seems that I managed to install QE 4.0.5 on a Mac Pro with 2 64-bit Intel Xeon processors. For the moment I have compiled a serial build (didn't try a parallel one yet) with gfortran and gcc 4.3.3 from the FINK package. Both pw.x and cp.x seem to run properly. Mac OSX (10.5) apparently defaults to 32-bits, and I experienced memory limitations also with other applications. So I wanted 64-bit compilation. Therefore I issued sudo ./configure --disable-parallel CC=gcc-4 CFLAGS="-m64" CPPFLAGS="-m64" LDFLAGS="-m64" FFLAGS="-m64" Note that I didn't set the "-m64" option in F90FLAGS, since this would result in compilation errors with iotk sources. Then, in order to have compatibility between lapack_atlas and dlamch at the 64-bit level I had to modify manually make.sys as follows FFLAGS_NOOPT = -O0 -m64 (I use lapack shipped with the QE package). Hope this helps and thanks for all the suggestions. Cristian -- ___________ Cristian Degli Esposti Boschi CNR, CNISM, Unita' di Ricerca di Bologna, c/o Dipartimento di Fisica, Universita' di Bologna viale Berti-Pichat, 6/2, 40127, Bologna, Italia tel. ++39 051 2095114 fax ++39 051 2095113 e-mail: cristian.degliesposti -AT- unibo.it web:http://www.df.unibo.it/fismat/theory ___
[Pw_forum] Bond-order-corrected (or BOLS) pseudopotentials
We would like to understand better the so-called bond order length strength (BOLS) correlation approach [Chang Q. Sun an co-workers, basic ref. is Prog. Solid. State Chem. 35, pag. 1 (2007)] to structures with under-coordinated atoms (surfaces, ribbons, rods, etc.) Sometimes it is stated that the BOLS approach is incorportated with DFT calculations, and in particular quoting from [J. Phys. Chem. C, 112, 18927 (2008)] "...Unfortunately, the presence and polarization of the nonbonding states are hardly described from the self-consistent optimization in DFT or tight-binding approximations at the present because of the limitation of interatomic potentials. Potentials suitable for the broken-bond-induced local strain and quantum trapping that are sometimes several folds deeper than the potential at the usual atomic site in the bulk and potentials suitable for the weak interaction between the nonbonding and the polarized antibonding states are highly desirable for computations." It seems to me that they would call for pseudopotentials corrected by the fact of going to represent atoms with low coordination or nonbonding states/electrons. Typically, in such a situation, the neighboring bonds tend to shrink and the bonding energies (in magnitude) tend to increase. Should one then consider a "BOLS-corrected" pseudpotential? Are you aware of problems of this kind, or is the question correctly posed at all? Thanks again for your time. Cristian
[Pw_forum] Bond-order-corrected (or BOLS) pseudopotentials
Sorry for forgetting it in my previous post; my affiliation (now) is: Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia
[Pw_forum] two conventions in postprocessing pp.x
Dear developers, in order to perform some quantitative calculations we would need to know precisely the conventions used in the postprocessing output of pp.x regarding the following points: plot_num = 1: is this the potential or the potential energy, that is, in atomic units is there a multiplicative factor e=sqrt(2) (electric charge) or not? iflag = 0: does the sperichal average include a normalization factor 1/(4 pi) or not? Alternatively, we may check the appropriate code: which one? Thanks a lot for your time. -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] compatibility between pseudopotential choice and input_dft token
Dear all, in these days I am doing some basic attempts using the implementation of van der Waals interactions within QE, taking for example carbon-based systems. My question is about the compatibility (both at the code level and even at the physical level, if the question is addressable in this forum) of the form chosen through the input_dft token on the one hand and of the actual pseudopotential file on the other hand. Any combination is possible? Meaningful? For example, can I safely set input_dft="vdw-df" and choose a LDA pseudo (say C.pz-vbc.UPF)? Or a GGA (say C.pbe-mt_fhi.UPF in my case)? The same with input_dft="vdw-df-c09" or "vdw-df2-c09" ? Thanks in advance for any help. Have a nice weekend. Cristian -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] compatibility between pseudopotential choice and input_dft token
>> Or a GGA (say C.pbe-mt_fhi.UPF in my case) > All the vdW corrections (semiempirical DFTD2 included, london=.true., > in &system) are well tested with a given kind of xc functional. This > is often discussed in the related papers. If you are not certain of > > the functional recipes, look into the funct.f90 file. But this is not > a hard limit. You may build your own functional, including vdW. I've > tested, with PBE pseudopotentials, an hybrid+vdW functional time ago > ('sla+pw+hse+vdw1'), with good results on molecule-surface > interactions. Thanks for the answers; now my question can be more specific: suppose I want to use a pseudopotential built constructed for PBE (say C.pbe-mt_fhi.UPF in my case). If I set input_dft = "vdw-df", that means "sla+pw+rpb+vdw1", as far as I can see the gradient correction on exchange is revPBE and not PBE (third index is 4), and also there is no gradient correction on correlation (fourth index is 0). Is this compatible with a PBE pseudopotential file that, if used without vdw-df specification, would report an exchange-correlation scheme "sla+pw+pbx+pbc" (1 4 3 4 0)? Or, should I set something like "sla+pw+pbx+pbc+vdw1" in order to be consistent with the PBE pseudopotential file? The same question may hold for Cooper's variant (c09x instead of rpb above) The question is on how the code works internally; at the beginning of the output there is a warning "Any further DFT definition will be discarded", and I wonder what it means exactly... Thanks again for your time. Cristian -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/
[Pw_forum] compatibility between pseudopotential choice and input_dft token
> it means that what is specified as "input_dft" is used, what is > written in the pseudopotential files is discarded. > P. Thanks again for the answers related to the choice of van der waals density functionals by means of the input_dft variable for pw.x To be precise, I guess that what is written in the pseudopotential file is discarded only as long as the form of the exchange-correlation functional is concerned. I mean, what is actually written in the pseudopotential file for a specific atomic element is of course used to build the ionic interactions (including core electrons). For example, I tested two different available pseudopotential files for carbon, namely C.pbe-mt_fhi.UPF and C.pz-vbc.UPF (both norm-conserving but built with different xc models, I guess...) and the rest of the input file being identical, including input_dft="vdw-df", and the resulting output are of course different meaning that something read from the UPF files makes the difference. So, input_dft="vdw-df" overrides the specification of the xc functional by imposing "sla+pw+rpb+vdw1" in both cases but at some point the two different files will show up differences due to the fact that they were built using different xc schemes; am I right? If so, sorry for the somehow silly question and thanks again. Cristian -- Cristian Degli Esposti Boschi CNR-IMM, Sezione di Bologna, via Gobetti, 101, 40129, Bologna, Italia tel. ++39 051 6399152, fax ++39 051 6399216 email: degliesposti -AT- bo.imm.cnr.it web: www.bo.imm.cnr.it/site/staff/personal_pages/degliesposti/