WIEN2k-Update information:

Current version: WIEN2k_11.1

Version 11.1 is a very important update. There are several bug fixes included and particular important changes are:

• new mixer (developed by L.Marks). Besides improved scf-convergence (switches MSEC3 and MSR1), it has a completely new mode: MSR1a (in case.inm). This will simultaneously update the density but also the atomic coordinates according to the forces on all atoms. It is a bit in the spirit of Car-Parrinello, i.e. mixer tries to keep close enough to the Born-Oppenheimer surface so that on average the forces are not complete nonsense and optimizes everything at once and not in the usual sequential way (optimize density for given coordinates, move atoms, restart the next scf-cycle,...). This scheme, which probably will be further improved in future versions, may save up to 70% of the required cpu-time to optimize the internal positions of a structure. Tests so far have shown that in most cases the scheme is quite stable (maybe even more stable that PORT/MSEC1) and reduces (sometimes only little) the cpu-time. However, we must admit that we also found cases where it diverged. For further reading see the UG (structure optimization) and Mixer_README_4.1.pdf(textbook-section).
• XMCD calculations have been added to the OPTIC package thanks to Lorenzo Pardini (loren.pard@gmail.com)
• TELNES3 by K.Jorissen and C.Hebert. It replaces the old TELNES package.
• At http://www.wien2k.at/reg_user/unsupported you can find various new or updated utilities like:
  • Wien2Wannier by Jan Kunes and Philipp Wissgott. It provides an interface between WIEN2k and Wannier90.
  • scf2forces by Evgeniya Kabliman (TU Vienna), an interface between WIEN2k and the PHONOPY code of Atsushi Togo.
  • ATAT@WIEN2k by M. Chakraborty, J. Spitaler, P. Puschnig and C. Ambrosch-Draxl. This package provides an interface to ATAT (cluster expansion and phase diagrams of alloys).
  • Cubic-elastic: by Morteza Jamal (updated version for elastic constants of cubic systems).
• interface to TRIQS (DMFT) calculations based on projective Wannier function construction, according to Aichhorn/Pourovskii/Vildosola/Georges/Biermann et al. For details, see Phys. Rev. B80, 085101 (2009), or http://www.ipht.cea.fr/triqs

To update from previous versions you can either download and install the complete package (highly recommended when upgrading from a rather old version) or download and update only individual packages (the ones which have been changed according to the list given below):

• SRC_w2web: just expand: gunzip SRC_w2web.tar.gz;tar -xvf SRC_w2web.tar
• SRC: expand and cp the respective shell-script to $WIENROOT: gunzip SRC.tar.gz;tar -xvf SRC.tar;cp SRC/CORRECTED_SCRIPT_lapw .
• For all other packages (Fortran programs) use the "update" option of siteconfig_lapw.

• SRC_w2web: Unfortunately SRC_w2web was not updated in WIEN2k_11. The corrected version supports now TELNES3 and includes several bug fixes ( plotting using gnuplot ) and enhancements.

• SRC: init_lapw (error handling for sgroup, symmetry, R0), init_phonon_lapw (fix for run_phonon script), initso_lapw (hints to avoid "NULL-MATRIX error"), make_inso_lapw (fix for identical atomnames); instgen_lapw, migrate_lapw (.lcore added); runeece_lapw (different default for xc-grid, which has some influence on the final results); run*_lapw (support for -qdmft and MSR1a structure and scf minimization); runsp_c_lapw (runs lapw2 only once); setrmt_lapw (additional switch -a Ti:1.95,C:1.7 for fixing RMTs at certain values); siteconfig_lapw (support for gfortran, ifort12, MKLROOT-variable), testconv_lapw, vec2old_lapw (support for rcp, better $SCRATCH support); vec2pratt_lapw, x_lapw (support for join_vectorfiles,telnes3,dmftproj,cif2struct, mixer,XMCD,detection of complex case, -alm* -fbz -telnes -qdmft switches, better writing EF into case.in1); *para_lapw (pwd error message fixed, allow for "." in hostnames, remove temporary files); dosplot2_lapw,dosplot,Curve,opticplot,rhoplot,specplot (compatibility with newer gnuplot), write_inq_lapw (for telnes3, creates case.inq from case.innes)
• SRC_cif2struct: set accurate "1/3" for hex systems
• SRC_clmcopy: support for format with 1 more significant digit
• SRC_fsgen:hex_fs_mesh.f (support for one more plane)
• SRC_joint: support for XMCD calculations added by Lorenzo Pardini, several new subroutines added
• SRC_kgen: check for plausible "nx,ny,nz" input
• SRC_lapw0: mBJ routines enhanced (heavy elements, strange tau values), sequential fftw-support added (to gain a (small) speedup, you have to use the fftw-routines from the mkl, not from the ordinary (self-compiled) fftw-sources. In order to be able to use them, you have to compile the corresponding mkl-interface. See the UG for some hints on that. In addition you have to add in the Makefile FOPT: -DFFTW2 and add the fftw-library -lfftw2xf for linking), modified xcpot-mesh inside spheres because of symmetry breaking --> may change slightly the final results
• SRC_lapw1: lopw.f: PW-selection changed to stricter criterium (can avoid ghostbands in some cases); SymmRot.f added, correct reading of vorb (fixing early 10.1 error for complex cases);
• SRC_lapw2:sequential fftw-support (see comments for lapw0), small speed-up (mpi, forces), support for DMFT calculations (Aichhorn/Georges/Biermann et al.)
• SRC_lapwdm: small fixes, better rotdef.f
• SRC_lapwso: correct reading of vorb (fixing early 10.1 error for complex cases); garadme.f: adding stupid write statement for fixing ifort12 bug
• SRC_lcore: corrected case.rsplcoreup/dn files for spin-polarized core-superposition
• SRC_lstart: writes MSR1 as default mixing option into case.inm
• SRC_mixer: major changes by L.Marks. MSR1, MSEC3 modes, MSR1a mode for simultaneous optimization of density and posititions. For a detailed description see the UG and the Mixer_README_4.1.pdf file.
• SRC_optic: support for XMCD calculations added by Lorenzo Pardini, see UG
• SRC_phonon: support for PHONON_5, which allows atom-names with up to 3 characters
• SRC_qtl: qtlmain.f fix for local rotations by Masao ARAI
• SRC_spaghetti: comprel modified for one case (LaRu2P2)
• SRC_symmetry: higher accuracy for detection of positions "near" 1.0; wrong multiplicity of an atomic site gives now an "ERROR".
• SRC_symmetso: fix for reading case.dmat/vorb (complex cases), support for triclinic lattic, higher accuracy of positions
• SRC_telnes3: new package by K.Jorissen, replaces SRC_telnes
• SRC_templates: case.in7 added
• SRC_tetra: small bug-fix
• SRC_trig: new program: join_vectorfiles to combine "parallel" vertor/energy files into a single file. (Contributed by P. Wissgott, TU Wien).
• SRC_vecpratt: various small fixes
• SRC_w2web: you can now delete a session AND its data

• SRC: various fixes and improvements for the following scripts: clean_lapw, clmextrapol_lapw (does not destroy clmsum file when case.rsp is missing), extract_aim, init_lapw (support for core-density superposition with .lcore file; WARNING for unphysical R0); initso_lapw (rewritten by Morteza Jamal, together with make_inso_lapw it lets you define various options when generating the required input file case.inso); lapw1_para; lapw2_para ( -qtl option works now also with parallel vectors on $SCRATCH); min_lapw (tests for case.rsp before it does a charge extrapolation, new options to support iter.diag.); opticpara_lapw (correct handling of .processes file); restore_lapw;
  all run*_lapw scripts:
  • by default EF is inserted into case.in1 and lapw1 will replace all default energy parameters ("0.3") by EF-0.2. When you want to keep the old behaviour, you must specify -in1old (and keep EF at 0.5Ry).
  • iterative diagonalization: -it will try to select iter.diag from the first iteration on; -it1 will recreate case.storeHinv files at first iteration; -it2 will do a full diagonalization, then create case.storeHinv in the 2nd cycle, and thus reach iter.diag speed only in the 3rd iteration.
  • -noHinv lets you switch to the iterative diagonalization without storing Hinv (maybe better in cases of slow I/O or difficult cases)
  • -vec2pratt use vec2pratt instead of vec2old for iterative diag. (L.D.Marks)
  • "touch .noHinv" removes case.storeHinv files and restarts iter.diag
  • Printing header into dayfile and case.scf (:LABEL)
  • case.in0_grr will initiate a second call to lapw0 using this input file. Necessary for modified Becke-Johnson potential or for decomposing E-xc into E-x and E-c.
  • the presence of .lcore file will remove case.clmcor and initiate a superposition (x dstart -lcore) of the core densitites creating case.clmsc. .lcore will be created automatically by init_lapw when you neglect core-leakage warnings. Note, since dstart is not parallelized, it can be slow for large cases.
  save_lapw; siteconfig_lapw (support for ifort 11.1, hints about ifort-linking and environment); testconv_lapw (-cc will not take the average :DIS, but the maximum for a particular atom, which is a much tighter criterium); vec2old_lapw (-local switch copies from $SCRATCH to the local directory); vec2pratt_lapw (by L.Marks, makes a Pratt-mixing of old and new vectors instead of a copy for iterative diagonalization); write_in1_lapw; x_lapw ("x paihess -copy" copies the files for usage in min_lapw, "x dstart -lcore" superposes core-densities, "x lapw0 -grr" calls lapw0 with case.in0_grr, "x lapw1 -it -noHinv" selects iter.diag. without Hinv, various other updates, "x lapw2" will insert EF into case.in1)
• SRC_w2web: "Tools": supports generation of .lcore, case.inm_vresp, case.in0_grr; "download" option fixed; support for new lstart option;
• SRC_2Doptimize: produces proper backslash character in scripts.
• SRC_cif2struct: fix for inaccurate 1/3,... in hexagonal SG.
• SRC_dipan: fix for more than 99 atoms
• SRC_dstart : support for superposition of core densities
• SRC_joint: small fix in band analysis
• SRC_lapw0 : support for "modified Becke-Johnson" potential to obtain proper band gaps; possible analysis of E-tot with contributions of E-kin and E-xc; total energy corrected for charged cells (should now be formally correct, but does not include ad hoc corrections due to periodic images); several input enhancements like a "print switch" to reduce output0 file or a PARAMETER "restrict_output" (in param.inc) to limit the number of case.output0XXX files for lapw0_mpi; rotdef.f "symmetrized"; W2kinit.F and W2kutils.c added (thanks to L.D.Marks; sets stacksize to unlimited); changes in mpi-version : "crash security added", "EFG-structure" and calls to MPI_ADDRESS or MPI_GET_ADDRESS removed (should work now with MPI 1 and 2);
• SRC_lapw1 : tries to read EF from case.in1 and modifies all default energy parameters (0.3) to EF-0.2; bug-fix for L=7 "cubic harmonics"; iterative diagonalizations : switch between storing or not Hinv added (activated by the presence of files at unit 200 and 201, see x_lapw); tighter test for approximate linear dependency when selecting a K-vector for the local orbitals (removes sporadic " choleski stops "); rotdef.f "symmetrized"; W2kinit.F and W2kutils.c added (thanks to L.D.Marks; sets stacksize to unlimited); changes in mpi-version: "crash security added"; PARAMETER "restrict_output" (in param.inc) to limit the number of case.output1_y_proc_XXX files for lapw1_mpi
• SRC_lapw2: bug fix in fourir.F: for some lattices (mainly centered lattices) it could have happened that not all non-zero Fourier-coefficients of the density have been calculated, but the latests (smallest) are missing;rotdef.f "symmetrized"; W2kinit.F and W2kutils.c added (thanks to L.D.Marks; sets stacksize to unlimited); changes in mpi-version: "crash security added",PARAMETER "restrict_output" (in modules.F) to limit the number of case.output2_y_proc_XXX files for lapw2_mpi
• SRC_lapwdm: speed-up for parallel run and only 1-kpoint per parallel job;
• SRC_lapwso: rotdef.f "symmetrized";
• SRC_lcore : bug fix for potential shift (outside RMT), which was the reason that sometimes a potential shift still did not help to put high-energy states as core; writes core density beyond RMT suitable for superposition with dstart; iprint switch added to reduce size of case.outputc
• SRC_lstart: WARNING if R0 is too big; default "NBAND" in case.in1 for iter.diag. increased; alternate selection of "core" states by "charge confinement" instead of energy (in particular for 4f semi-core states of 5d elements), case.outputst contains "readable" information on core state selection and suggestions to touch .lcore for core density superposition (see run_lapw description); first line in case.in1_st has been changed.
• SRC_mini: proper format option of case.scf_mini;
• SRC_mixer: fixes sevire bug for case.dmatdn (case.dmatdn was never mixed but taken completely from the new output. There is no problem when your calculations converged); support for E-kin output; small fixes for spin-pol. case and accuracy
• SRC_nn: prints name of newly generated case.struct_nn file
• SRC_optic: speed improvements, optional input nbvalmax (number of full occupied bands) to save time;
• SRC_orb: small fix
• SRC_qtl: small fix for more than 27 atoms
• SRC_supercell: CXY structures can produce supercells with P or CXY lattice
• SRC_symmetso: fixes for monoclinic angle and centered cells (similar to SRC_symmetry)
• SRC_tetra: can produce a DOS with a single k-point (for big cells with more than 100 atoms) using Gaussian or Lorentzian (or both) broadening (see case.int description)
• SRC_trig/xyz2struct.f: several improvements (rounding errors for H lattice)
• SRC_vecpratt: new package contributed by L.D.Marks. Can alternatively be used in iterative diagonalization. It will Pratt-mix two old vectors instead of just copying. (activated with the -vec2pratt switch in run_lapw)
• SRC_lib/lapack_lapw: new subroutines included
• The "unsupported software" section became new packages, in particular for phonons, wien2wannier, critic, and elastic constants.

• SRC: various fixes and improvements for the following scripts: ana2D_lapw, clean_lapw, dosplot2_lapw, init_lapw, init_phonon_lapw, instgen_lapw, lapw0para_lapw, lapw1para_lapw, lapw2para_lapw, lapwsopara_lapw, min_lapw, restore_lapw, runfsm_lapw, runsp_c_lapw, runsp_lapw, save_lapw, siteconfig_lapw, testpara_lapw, x_lapw.
  setrmt_lapw (slightly modified RMT settings), write_in1_lapw has a new option [-ef]. When specified, it will only update the default energy-parameters 0.3 to EF-0.2. All scf scripts (run_lapw, ....) take a new switch -in1ef, which activates the new option.
  I recommend using -in1ef instead of -in1new N.
• SRC_w2web: band.pl (xmgrace file download), dos.pl (download of ASCII-DOS files), executor.pl (instgen_lapw, lstart-PBEsol, -in1ef switch), scf*.pl (-in1ef switch), extract-dos-cases_lapw, initlapw.pl (case.inm_st removed, AFM messages/logic), min.pl (better pairhess support), optic.pl (ASCII-download), rho.pl (case.rho download), single.pl (instgen support), xspec.pl (ASCII-download)
• SRC_2Doptimize: setup2D.f
• SRC_aim: aim.f, charge.f, dtdtylm.f, dtylm.f, follown.f, moduls.frc,outin.frc ,rsurf.f,sphere.f,sternb.frc,surf.f,
• SRC_arrows: allows now to display differences between two structures or "forces on the atoms" from a single calculation.
• SRC_eosfit: prints now also the "cubic" lattice parameter (not just the equillibrium volume).
• SRC_dipan: calculates the off-site dipolar contributions to B_dip and E_anisotropy. Thanks to P.Novak.
• SRC_fsgen: fcc_fs_mesh.f (new plane added)
• SRC_kgen: removed printout of "wrong" symmetry operations
• SRC_kram: fix for (zero) broadening
• SRC_lapw0: various changes in several files, support for new DFT functionals (PBEsol)
• SRC_lapw1: coors.f (dimensioning with min(nume,nv+nlo)), new and further improved iterative diagonalization
• SRC_lapw2: deter.f fixed, scf-file contains ZZ and name of atom;
• SRC_lstart: larger default for number of bands in iter.diag., PBEsol support
• SRC_mini: small adaptions to PORT
• SRC_mixer: various improvements and speedup
• SRC_nn: print "bond-valences", optional "accuracy limit" for equivalent atoms
• SRC_pairhess: various improvements for better starting Hessian; includes new programs struct2cif, struct2xyz and eigenhess (see UG). Thanks to L.D.Marks
• SRC_phonon: small fix
• SRC_sumpara: small fix
• SRC_supercell: R lattices can now be transfered into H supercells
• SRC_symmetso: various fixes (thanks to P.Novak)
• SRC_trig: improved xyz2struct.f
• SRC_tetra: support for RXES spectra (k-dependent weights)

• SRC: save_lapw, restore_lapw (adding vresp-files); instgen_lapw (massive change: allows to select spin-states for each atom); init_lapw (support for new instgen_lapw); lapw0para_lapw (); lapw1para_lapw (small fix for certain hostnames); addjoint-updn_lapw (support for multiple columns); siteconfig_lapw (support for pathscale compiler, rsh/ssh selection for vec2old, fftw_mpi-support); configure_int_lapw (thanks to Morteza Jamal; easy input generator for case.int); clean_lapw (-r switch cleans recursively all subdirectories); userconfig_lapw (gets shell from getent command); x_lapw (fix for $argv1 problem, support for arrows, modifications for qtl), init_phonon_lapw (support for batch-mode of init_lapw and initialization of every 2nd case for pos/neg displacements); ana2D_lapw and set2D_lapw (for 2D_optimize)
• SRC_2Doptimize: thanks to Morteza Jamal; package for 2D structural optimization (Volume and c/a; i.e. tetragonal and hexagonal structures) with convenient analysis
• SRC_arrows: thanks to E.Kabliman, interface program to xcrysden. It allows to draw arrows which indicate the atomic movements during a structure optimization.
• SRC_lapw0: lapw0_mpi includes fftw_mpi parallelization (needs additional fftw package, but this reduces drastically the memory requirements (thanks to R.Laskowski). Various other optimizations and fixes also for the sequential version. New hybrid-functionals implemented (with WC and PBEsol).
• SRC_lapw1: improving memory management,
• SRC_lapw2: unformatted *tmp files, fix for ALM option, printing :GAP for insulator
• SRC_lapwdm: small fix
• SRC_lib: blas_lapw updated for generic version (thanks to M.Divis)
• SRC_lstart: changes for some defaults in case.in1 and case.inq
• SRC_mixer: reduction of memory; new option MSEC2 mixes higher LM terms only with pratt-mixing for cases where memory or disk-space of broyden files is a problem.
• SRC_optic: new option to analyse contributions from different atoms to the optical tansitions (see UG).
• SRC_orb: small fix
• SRC_phonon: phonon2wien2k_sym.f: small fix for hexagonal lattice
• SRC_qtl: several fixes (spin-orbit, f-splitting, rotations,...) by P.Novak, J.Rusz and PB. Can now be used for TELNES (option 88,99) without "full k-mesh".
• SRC_spacegroup: runs now for monoclinc spacegroups AND angles=90 degree. However, please note that this is potentially dangerous, since for certain monoclinic spacegroups the "monoclinic axis" is then not defined uniquely and we make a certain assumption (usually angle gamma.ne.90) for this.
• SRC_tetra: small fix
• SRC_trig: xyz2struct.f: various improvements to enhance flexibility (see UG).
• SRC_w2web: support for new features (new instgen_lapw in initlapw.pl; executor.pl, conf_int_lapw in dos.pl, band.pl, edit,pl, structend.pl)

• SRC: addjoint-updn_lapw added (for spin-polarized optics); lapw0_para, opticpara_lapw (tail +2 problem fixed), x_lapw; userconfig_lapw updated; run_lapw (fixed inM-zero force test); runfsm_lapw (rewritten from runsp, supports now LDA+U); lapw0para_lapw updated;
• SRC_dstart: small fix
• SRC_fsgen: update for cxz_mon_fs_mesh.f
• SRC_joint: arbdos.f (small fix)
• SRC_lapw0: sogga.f added
• SRC_lapw1: BLOCKSIZE automatically adapted (L.Marks, may speedup mpi-version)
• SRC_lapw2: fix for mpi-version, modus ALM creates file case.almblm and case.radwf
• SRC_mini: various small improvements
• SRC_nn: small fix
• SRC_orb: init.f (small fix)
• SRC_pairhess: various fixes, obeys constrained positions in case.inM
• SRC_qtl: various fixes (dimensioning problems), SO analysis works also for non-spinpol case, normal basis works for SO calculation
• SRC_spaghetti: spag.f (better explanations when spaghetti stops because of inconsistent irrep-files)
• SRC_trig: add_columns utitlity program added (for addjoint-updn_lapw)
• SRC_w2web: bugfix for optimize (initial_struct); optics (spin-polarized case), new case ($W2WEB_CASE_BASEDIR and new_case_name as directory suggested); analysis (MMT-label), email notification

• SRC: various fixes and improvements in analyse_lapw,init_lapw,lapw1para_lapw,lapw2para_lapw,lapwdmpara_lapw,min_lapw (calls pairhess when case.inM is not present), runeece_lapw,run_lapw (test of force convergence only for "active" atoms and after ec,cc is fulfilled),scf_monitor_lapw, siteconfig_lapw (support for ifort10+mkl10), x_lapw; new qsub-*_lapw template-files for support of queuing-systems in w2web (see UG)
• SRC_cif2struct: (accepts "X,Y,Z")
• SRC_dstart: consistent K-list with lapw2 for some cases, consistent gpoint.f,in0_std defaults
• SRC_fsgen: new template for monoclinic CXZ lattice
• SRC_kgen: adds inversion automatically except when you call "x kgen -so"
• SRC_lapw0: param.inc (new lmax2x), xcpot1.f (accuracy increase, support for 2 gpoint-grids (pos/neg LUSE in case.in0); new iex=19 (PBEsol),iex=20 (AM05); improved accuracy for drho.f, dylm.f, grans.f (theta=0); vxclm2.f (accuracy for small rho, speed, PBEsol, AM05 functionals; lapw0.f (fix and improvements for EECE/Hybrid calculations, EECE is now mpi-enabled, support for TPSSh)
• SRC_lapw1: WARN for NUME-limitations, fix for VORB with multimple l; fix for ifort10+mkl10
• SRC_lapw2: Forces for LDA+U and hybrids implemented, improved EF, fixes for EECE+VRESP
• SRC_lstart: better warnings for core-leakage, writes case.inq_st
• SRC_mini: symmetrization of equivalent positions for H,R symmetry, better constrains
• SRC_mixer: automatic mixing for vresp=pratt1.0; new -eece switch (vorb-mixing)
• SRC_orb: fix for eece (error when LM list was different for different atoms), corrupted vorb_old file now accepted,
• SRC_qtl: enhanced version with much more userfriendly input. Provides partial charges for f-states, SO-coupling (p1/2,3/2) or different local rotations (thanks to P.Novak)
• SRC_spaghetti: cartco.f (monoclinic CXZ lattice)
• SRC_structeditor: various small fixes
• SRC_tetra: support for more columns in case.qtl (f-state splitting)
• SRC_w2web: support for new QTL-program

• VERSION_08.1: 14.12.2007
• SRC: new scripts dosplot2_lapw, Curve_lapw (see UG, thanks to Morteza Jamal). Fixes for lapw0/1/2/dmpara; min_lapw (charge extrapolation becomes default). All run*_lapw scripts have set riter to 99 (will not remove broyden files); runfsm_lapw (fix for iterative diagonalization); runsp_lapw (fix for -eece); siteconfig_lapw, vec2old_lapw, x_lapw (x dstart -super); save_lapw (removes all *.broyd* files)
• SRC_broadening: small fix
• SRC_clmaddsub: fix for spin-polarization and different k-vector lists
• SRC_clmcopy: fix for rho_k
• SRC_dstart: fix for R-lattice (K-vector list); enhancementfactor=1 for large GMAX; default case.in0_std with IFFT 0 0 0 2. (or 1. for H-systems)
• SRC_initxspec: "100 atoms" fix.
• SRC_kgen: new optional input: specify 3 mesh-divisions (instead of total number of k-points)
• SRC_kram: fix for broadening (Gamma must not be smaller than delta-E)
• SRC_lapw0: several small changes (small fixes, internally or for some test-potential)
• SRC_lapw1: changes to scf1/output1 file (new labels :EIG :E...); improvements for iterative diagonalization (removal of linear dependent terms, optional input parameter in case.in1, fixed possible change of number of LOs); fix for parallel complex iterative diagonalization (was completely broken in previous version); fix in sphbes/make_albl.F for low L values (LDA+U calculations); fix when case.vns file is not present
• SRC_lapw2: fermi.F: correction for degenerate states at EF; print occupancy next to bandranges
• SRC_lapwdm: fix for orbital (wrong sign for M=(010)) and dipolar HFF
• SRC_lstart: small fixes; default mixing 0.2 in case.inm.
• SRC_mixer: new version (thanks to L.Marks and R.Luke), should be even faster than the previously improved version.
• SRC_nn: samll fix for supercell calculations of single atoms in a box.
• SRC_optimize: rewriting of "optimize.job"; support for extrapol_lapw.
• SRC_orb: small fix for orbital polarization.
• SRC_pairhess: fix for F and B centered lattices (old version did not obey symmetry constrains, leading to changes in symmetry-constraint positions.
• SRC_sumpara: small fix
• SRC_telnes2: small fix
• SRC_trig: new programs fmax.f and Pinver.f (work together with dosplot2_lapw; thanks to Morteza Jamal) and struct2cif.f (thanks to F. Boucher)
• SRC_w2web: structgen will now correctly accept "a" or "0" for cubic lattice constants; support for the new kgen option; minimization: adaption for charge-extrapolation (default); new dos-interface (based on dosplot2)

• SRC: analyse_lapw (fix);new bashtime2csh.pl_lapw and clmextrapol_lapw (for charge density extrapolation during min_lapw) scripts; init_lapw (support for iterative diag (in1) and new_super.clmsum); initso_lapw; lapw*para_lapw (better support for bash, support for new mpi-versions); migrate_lapw; min_lapw (support for charge extrapolation and iterative diag, -ex, -it); opticpara_lapw ($scratch removed); run*_lapw (-it switch changed, vec2old works also for $SCRATCH, *broyd* files NOT removed after 20 cycles); save_lapw/restore_lapw (saves also case.vsp*); siteconfig_lapw (support for gfortran, seems to be a very good compiler, use it together with GOTOlib); vec2old_lapw (for iterative diag); write_in1_lapw; x_lapw (support for new/modified options (-emin X, -all X Y for lapw2), clmaddsub)
• SRC_w2web: various changes, enhancements, fixes: structeditor supports xyz and txt files, does not require b+c lattice parameter in cubic cases, support for new options (scf and min), support for SO in various tasks, support for batch-option in initialization, optimize allows to chose struct file, restore gives list of possible files)
• SRC_afminput: (fix)
• SRC_aim: (fixes, d1mach, )
• SRC_broadening : valencebroadening.f
• SRC_clmaddsub: new program to add and subtract two clmsum files from another one (for charge density extrapolation)
• SRC_dstart: major rewriting (using "xcpot1-philosophy" inside the spheres), significant speed-up (5 min vs. 20 h !!!) and reduced memory for cases with many atoms/cell and low symmetry
• SRC_joint: fixes
• SRC_lapw0: support for luse (determine angular fitting mesh), memory and speed improvements, speedup in nonspinpolarized case
• SRC_lapw1: completely rewritten routines for mpi performance ( hamilt, horb and hns modules,...), new iterative diagonalization (gives now accurate results and is much faster than full diag.), mpi version supports now also orbital potentials (LDA+U).
• SRC_lapw2: completely rewritten routines for mpi performance and reduced I/O, additional commandline argument (vectorsplit) for mpi-version with memory reduction, vresp and help-files only on demand,
• SRC_lapwdm: fix in xoper.f routine
• SRC_lib/lapack_lapw: new routines added
• SRC_lstart: support for creation of modified input files (inm,in1)
• SRC_mixer: new MSEC1 option contributed by L.Marks and R.Luke. In most cases this gives a SIGNIFICANT improvement, both, reducing the number of iterations and in stability. Makes old mixer options rather obsolete.
• SRC_spaghetti: improved ps file (for character plots, where gv display was often wrong)
• SRC_sumpara: small fix for dmatscf.f, sumpara.f
• SRC_telnes2: latgen.f: sevire bugfix (conversion of angles to rad was done twice)
• SRC_trig: xyz2struct.f: improvements

• SRC: siteconfig_lapw (g95 setup with "W" fixed), runsp_lapw (fix for -in1new and -eece)
• SRC_dstart/dstart.f: (rhok initialized with zero)
• SRC_lapw0: eramps.f (integer do-loop)
• SRC_lapwdm: couplx.f (small fix for MS)
• SRC_lib: lapack_lapw library includes missing routines for SRC_mini
• SRC_elast: anaelast.f (format) and genetempl.f (severe bug fix for MULT > 1)
• VERSION_07.1a: 31.01.2007
• SRC: init_lapw (support for "batch mode" for "trivial" cases). migrate_lapw, lapw1para_lapw + lapw2para_lapw + fixerror_lapw + lapwsopara_lapw + lapwdmpara_lapw ("NFS-bug fixes"), lapw0para_lapw, save_lapw + restore_lapw (case.scf2 files for -in1new switch), write_in1_lapw, run_lapw, runsp_lapw + runeece_lapw (for hybrid methods, see UG), clean_lapw, dosplot_lapw, opticpara_lapw, runafm_lapw (+ struct_afm_check to enforce AFM structure), x_lapw, siteconfig_lapw (support for g95), userconfig_lapw
• SRC_w2web: navig.pl, struct.pl, executor.pl, scf.pl, ...
• SRC_aim: d1mach.c
• SRC_cif2struct: f90 cleanup
• SRC_clmcopy: new program struct_afm_check.f added
• SRC_dstart: -fft switch to stop after in0_st creation;
• SRC_elast: anaelast.f: fix for an error if "MULT.ne.1"
• SRC_irrep: f90 cleanup
• SRC_joint: several small fixes
• SRC_kgen: bigger default dimensions, new format for case.klisti (the old format is still working, but you MUST update lapw1 to be compatible with both, the old and new format of case.klist).
• SRC_lapw0: several major changes: speed-up of rean0.F in parallel version; WC-potential is now option 11; fix for various bugs for charged cells and external electric field (Thanks to L.Marks), support for "exact exchange+hybrid functionals for correlated electrons". This means that on-site Hartree-Fock inside the spheres can be added (without much extra cost). The results will be "similar" to LDA+U and this method is applicable ONLY to localized electrons (It will NOT improve the band gap in semiconductors like Si).
• SRC_lapw1: atpar.f: trap when Alm=1; inilpw.f (support for alternate klist file format); inilpw.f: support for old and new klist-format
• SRC_lapw2: fermi.f (insulator for magnetic case fixed); :SUM written with more digits (may change :ENE; which was affected by rounding errors in parallel calculations.)
• SRC_lapw5: rotdef5.f
• SRC_lapwdm: significant improvement of the calculations of orbital or dipolar hyperfine fields by P.Novak. See Pavels notes in http://www.wien2k.at/reg_users/textbook/Bhf_3.ps
• SRC_lcore: core eigenvalues writted with more digits (may change :ENE; which was affected by rounding errors when summing over many core electrons.)
• SRC_lstart: WC potential added as option (PW91 removed)
• SRC_mixer: rean0a.f speedup
• SRC_optic: sph-UP.frc (support for more than 999 eigenvalues)
• SRC_optimize: fix vor variation of gamma
• SRC_orb: updates for Hybrid functionals
• SRC_pairhess: small fixes, f90 cleanup
• SRC_structeditor: small fixes
• SRC_sumpara: f90 cleanup
• SRC_symmetry: GMAX new default = 12 (20 for Hydrogen)
• SRC_symmetso: f90 cleanup
• SRC_templates: case.ineece
• SRC_tetra: allows more than 21 cases
• VERSION_06.4: 24.08.2006
• SRC: eplot_lapw (-f switch); grepline_lapw (-s for silent); init_lapw (support for new case.in0 from dstart), lapw1para_lapw, lapw2para_lapw (NFS-bug fix),parabol_fit_lapw (for more dim. structure optimization and new optimize options); run*_lapw (40 default iterations, better support for multiple convergence criteria); siteconfig_lapw (SUN,linuxif8 options), x_lapw (eosfit6, pairhess, constraints for mini, xmgr file for spaghetti, in0_std)
• SRC_w2web: initlapw.pl (case.in0_std after dstart); struct.pl (format change), min.pl (support for automatic pairhess call + inM file)
• SRC_*: identical rotdef.f in all programs, or renaming to rotdef1.f; corresponding changes in latgen.f,... when necessary; cputim.c and cputim_generic.c changed in all programs; struct-file format for symop and positions in all programs which read/write case.struct (8 digits after the comma)
• SRC_broadening: various fixes by Kevin Jorissen
• SRC_cif2struct: support for *.cif / *.CIF
• SRC_dstart: GMIN warning; FFT-mesh calculation and generation of case.in0_std
• SRC_eosfit6: new program (see UG), performs least squares fit for multi-dimensional lattice parameter optimization (parabolic approximation)
• SRC_joint: small bugfix
• SRC_lapw0: FFT-grid parameter + scaling read from case.in0 (optionally). It turns out, that for certain cases (lattic parameter optimization with GGA) a constant AND dense FFT mesh is very important for numerical stability. We have thus reduced the default GMAX, but introduced a fixed and finer FFT mesh (important only for E-XC in the interstital; increases slightly the xcpot3 time); support for the new Wu-Cohen GGA, which gives better results for equillibrium lattice parameters than PBE in many cases (except group-2 elements, 3d metals, molecules)
• SRC_lstart: radial mesh warning changed, f90 cleanup
• SRC_mini: linear constains of several positions (thanks to B.Yanchitsky), see $WIENROOT/SRC_templates/case.constraint
• SRC_optimize: new option 8 (vary C/A + Volume); small fixes
• SRC_pairhess: new program by J.Rondinelli, Bin Deng and L.Marks (http://xxx.lanl.gov/abs/physics/0608160). Creates an approximate hessian matrix (in .minpair) for structure minimization using the PORT option, reduces in many cases the number of geometry steps significantly. Usage (see UG): x pairhess; cp case.inM_st case.inM ; cp .minpair .min_hess ; cp .minpair .minrestart; min_lapw
• SRC_sgroup: code cleanup
• SRC_spaghetti: generates now case.bands.agr, which can be directly opened with xmgrace (thanks to juergen.spitaler@mu-leoben.at)
• SRC_symmetry: monoclinic angle gamma automatically constraint to 88-92 degree when finding symops; default-GMAX reduced to 12.0
• SRC_symmetso/class.f: redefinition of ISPLIT
• SRC_telnes2: various small fixes
• VERSION_06.3: 18.05.2006
• SRC: instgen_lapw,lapw1para_lapw, lapw2para_lapw (NFS-bug),migrate_lapw, siteconfig_lapw, x_lapw
• SRC_w2web: struct.pl (inst-file), executor.pl (supercell), initlapw.pl (case.inst updated with changes of case.struct), telnes2.pl (hint for expert-mode)
• SRC_aim: interst.frc (fixes temporary error made in previous distribution)
• SRC_clmcopy: Support for more than 999999 PW
• SRC_irrep: f90 cleanup
• SRC_joint: memory reduction (thanks to F.Boucher)
• SRC_kgen: small output adjustment
• SRC_lapw0: least squares fit removed, rean0 speedup (thanks to L.Marks), fft-params printed to scf0
• SRC_lapw2: fermi-energy with TETRA for insulators fixed (nonmagnetic only), real*16 removed
• SRC_lapw3: f90 cleanup
• SRC_lapw7: f90 cleanup
• SRC_lcore: f90 cleanup
• SRC_lstart: f90 cleanup, adjustment of EMIN/EMAX (case.in1,in2) to E-core, core-leakage warning
• SRC_mini: format fixes
• SRC_mixer: support for "open-core 5f", "big"-check for large diff on one atom
• SRC_nn: small fix for only one nn-atom
• SRC_spaghetti: f90 cleanup
• SRC_supercell: f90 cleanup
• SRC_symmetry: testing accuracy of positions increased by one digit
• SRC_telnes2: f90 cleanup
• SRC_templates: adjustments to case.injoint and case.inop
• VERSION_06.2: 9.03.2006
• Support for more than 999999 PW coefficints in clmsum/vns,... This required changes in SRC_dstart, lapw0, lapw1, lapw2, lapw3, lapw5, mixer, sumpara, symmetso, clminter. Old files are still compatible.
• SRC_lapw5: Severe bug fix (introduced in WIEN2k_06.1) in outin.frc. The density in the interstital region is wrong in WIEN2k_06.1.
• SRC_mini: fix for the NEW1 option (introduced in WIEN2k_06.1). It changed positions even when restricted by 0.0 in case.inM
• SRC: lapw0para_lapw,lapw1para_lapw,lapw2para_lapw: removing of some "parallel" files (vsp/vns/klist_*/weigh*) to work against the "NFS-bug".
• SRC_optimize: support for full optimization of lattice parameters (tetragonal, hexagonal, orthorhombic and monoclinic structure)
• VERSION_06.1a: 6.03.2006
• SRC: siteconfig_lapw (-i-static removed), userconfig_lapw (STRUCTEDIT_PATH corrected),
• SRC_w2web/libs/struct.pl: structeditor fix (wrong conversion to Ang when units are not specified in case.struct)
• VERSION_06.1: 23.02.2006
• SRC: law0para_lapw, x_lapw (status check added); siteconfig_lapw (LC_NUMERIC, lib-search), userconfig_lapw (environment variables); opticpara_lapw (spinpolarization fixed); run*_lapw and lapwpara_lapw (rm case.vsp/vsn/weigh* in order to fix a possible NFS-bug on some systems)
• SRC_aim: bader surface and charge integration rewritten by L.Marks. Significant speedup (up to a factor of 10)
• SRC_cif2struct: small fix
• SRC_dstart: small fix (CXZ bug)
• SRC_eosfit: format changes
• SRC_lapw0: fix for cubic LM > 6
• SRC_lapw1: cbcomb.f: small fix for L=8 component; proper (smaller) allocations for mpi-version leading to significant less memory;
• SRC_lapw2: QTL-B warnings changed (produces :WARN), fix for EF on insulators, mermin functional, other small changes
• SRC_lapw3: fix for cubic LM > 6
• SRC_lapw5: speedup, fix for cubic LM > 6
• SRC_lapwdm: small changes
• SRC_mini: new option NEW1 with adaptive steps
• SRC_mixer: better support when the K-list changes due to modified lattice parameters, dmat dimensions fixed; warnings in mixer changed
• SRC_nn: significant update. May lead to longer runtime, but (hopefully) more stable for extremely different lat.param (surfaces)
• SRC_spaghetti: format change in ps file
• SRC_sumpara: small changes
• SRC_supercell: support for optional shift of all atoms by a constant
• SRC_symmetry: check (and fix) that all positions are within 0 and 0.99999999
• SRC_telnes2: insld.f: read(18 changed
• SRC_templates: case.inaim and case.inM updated
• SRC_trig: plane.f: support for Hex lattice
• SRC_structeditor: new tools based on octave (mathlab) scripts and opendx to manipulate and view structures, and create arbitrary supercells and surfaces (written by R.Laskowski)
• VERSION_05.6: 20.10.2005
• SRC: eplot_lapw (support for Birch-Murnaghan);migrate_lapw (optional gzip,...); cancel_lapw; run*_lapw (iterative diagonalization interchanged with full diagonalization every Xth scf-cycle: -it X); opticpara_lapw; siteconfig_lapw, x_lapw
• SRC_w2web: small fixes, support for "-it X"
• SRC_lapw1: coors.f, tapewf.F, modules.F: fix of the "famous" monoclinic CXZ-bug (thanks to Georg Madsen !!) ; iterative diagonalization performs only one step (NUMIT=1 in secr5.F)
• SRC_broadening: fix for ABSORB and EMIS options
• SRC_eosfit: support for Birch-Murnaghan equation of states (thanks to S.Cottenier)
• SRC_lapw5: ylm.f replaced by numerical more stable version
• SRC_lapwso: speed (up to factor 3 without RLOs) and memory improvements for large cases
• SRC_telnes2: small fix in debug.f
• SRC_symmetry: small fix for triclinic lattices
• VERSION_05.5: 15.7.2005
• SRC: various small fixes or adaptations in opticpara_lapw (fix for new filenames), opticplot_lapw (support for more files); runsp_lapw, save_lapw (vorb* also saved), scfmonitor_lapw, setrmt_lapw, write_in1_lapw (search for states below -3 Ry), x_lapw
• SRC_w2web: small fixes, support for telnes2 (J.Luitz)
• SRC_broadening: new program devoted by K. Jorissen and C.Hebert for telnes.2 broadening
• SRC_dstart: somm2.f (double precission variables)
• SRC_lapw0: preperation for future options
• SRC_lapw1: support for "x lapw1 -nmat_only" (calculates only the matrix size and stops quickly)
• SRC_lapw2: QTL-B warnings changed (gives info on atom and l), preperation for future options
• SRC_lapwdm: small fixes
• SRC_lcore: replacement of various single precision variables
• SRC_orb: preperation for future options
• SRC_sumpara: support for the new QTL-B warnings
• SRC_symmetso: important bug fix for some symmetries (eg. hexagonal with M=100 direction). Previous version gave inconsistency with lapwdm and stopped.
• SRC_telnes2: replacement and greatly improved new version of the telnes program for the calculation of EELS spactroscopy. Developed by K. Jorissen and C.Hebert.
• SRC_templates: case.inM (R-trust), case.injoint, case.innes (adapted to telnes2)
• VERSION_05.4: 25.5.2005
• SRC: instgen_lapw can handle also "D" (Deuterium), fix to accept case.sturct_nn; usersguide update and new/modified chapters on TiO2-structure optimization and supercells; analyse_phonon: format fix; setrmt fixed; x_lapw: lstart -sigma
• SRC_w2web: Several changes, in particular: simplification of STRUCTGEN (only necessary fields are displayed), support for: -orb and -so in bandstructure, -sigma in rho; init (fix to accept nn-struct file); new options -band, -dm; Links to Bilbao Cryst.Server for spacegroups and Brillouinzones, support for element "D"
• SRC_eosfit: Corrected printout of Murnagan equation
• SRC_lapw0: small fix for workfunction
• SRC_lapw1: support for "x lapw1 -nmat_only", which prints the matrixsize (for the first k-point) to case.nmat_only (allows a quick estimate of needed ressources)
• SRC_lapw2: l2main.F, efgsplit.f, Makefile: Changes for the "EFG" option. Works now for all symmetries and gives p,d,f contributions. Thank's to G.Madsen!
• SRC_lapw3: speedup in various subroutines. Thanks to L.Marks!
• SRC_lcore: fix for He-atom case.
• SRC_mixer: rescaling of pw-mixing by number of atoms/cell
• SRC_spaghetti: small fix for simultanious "charactor-plot" of several atoms
• SRC_symmetry: small fix to prevent symmetry from hanging in an infinite loop for triclinic symmetry
• SRC_templates: case.injoint: has now a default of 999 for upper band index (so that the default input does not give "zero" epsilon2 for larger cases)
• VERSION_05.3: 3.4.2005
• SRC: important fix for min_lapw (a force of -xxx. was tuncated to +xxx.), small fixes for lapw0para, lapw1para (for "strange hostnames"), runfsm_lapw (complex parallel case), x_lapw (qtl uses scratch)
• SRC_w2web: small fix in structgen.pl
• SRC_lapw1: modules.F ( important correction for the complex mpi version)
• SRC_lcore: many routines changed to get full double precission accuracy. Note: your total energies may change slightly after this fix
• SRC_mini: important fix for NEWT option (the damping was nonsense for 2/3 of the atoms, so NEWT did a very bad job and oszillates a lot).
• SRC_mixer: small fix to allow format for even bigger forces, writes core-density inegrals
• SRC_qtl: small fix in write format
• VERSION_05.2: 25.2.2005
• SRC: init_lapw generates case.inst "on the fly"; small fixes for x_lapw, setrmt_lapw, siteconfig_lapw, opticpara_lapw, lapw0para_lapw
• SRC_aim: small fix
• SRC_cif2struct: small fix in Makefile.orig
• SRC_lapw0: small fix for mpi version
• SRC_lapw2: fix for Makefile.orig (sortag.o) and psplit.f
• SRC_mini: New optional input to determine "R-trusted" for first geometry step.
• SRC_spacegroup: small fix in Makefile.orig
• SRC_sumpara: dmatscf.f fixed
• SRC_telnes: inilpw.f stops on error
• VERSION_05.1: 30.1.2005
• SRC: setrmt_lapw (automatic selection of RMT values; based on a script by G.Madsen); init_lapw (includes new setrmt_lapw); migrate_lapw (moves a "case-directory" to/from a remote host); run*_lapw (support for multiple convergence criteria -ec -cc -fc); lapwdm_para; x_lapw; lapw1/2para_lapw ($WIENROOT inserted)
• SRC_w2web: support for setrmt_lapw; DISPLAY not set automatically anymore; adapted options for scf and min; analysis
• SRC_dstart: dramatic speedup, in particular of recpr.f (thanks to L.Marks)
• SRC_elast/anaelast.f: typo in output fixed
• SRC_joint: significant speedup (thanks to F.Boucher)
• SRC_lapw0: significant speedup in various parts (thanks to L.Marks)
• SRC_lapw2: recpr.f speedup; band-limits around EF to scf file; Ghostband check with stop
• SRC_lcore/fcore.f: small fix for L=10
• SRC_mixer/mixer.f: check for proper force symmetrization and more significant digits for :FGL (important for structural optimization with HEX symmetry, where a symmetry break was possible)
• SRC_sumpara/scfsum.f (adaption for QTL-B printout), sumpara.f
• SRC_supercell/supercell.f: small fix
• SRC_symmetry/pglsym.f,symmetry.f: small fix for triclinic lattic: NOFILE as default into case.in2
• 05.2004-12.2004: "List of changes in version 4"
• 01.2003-05.2004: "List of changes in version 3"
• 06.2002-01.2003: "List of changes in version 2"
• 12.2001-06.2002: "List of changes in version 1"

When you have a version prior to WIEN2k_03, you must install the complete package. Once you have WIEN2k_03, you can also update individual packages

Please note, from WIEN2k_03 on we support more than 99 inequivalent atoms/cell. For this purpose it must use a new format for case.struct, case.clm*, case.vsp*, case.vns* and case.in2* files. In addition, the output-format of case.scf had also to be adapted (3 digits for atom labels like :MMI001)

If you are upgrading from an earlier version (which is highly recommended, since we can support only the new version), you can convert all the files mentioned above using a procedure "100atomfix_lapw". (Conversion is done for the whole subdirectory tree and may take several hours when you have many large cases).

©2001 by P. Blaha and K. Schwarz