| 查看: 1142 | 回复: 0 | |||
wangyikeco木虫 (正式写手)
newbie
|
[交流]
WIEN2k_21.1was released on 14. April 2021
|
|
The current version is WIEN2k_21.1 and was released on 14. April 2021. Version 21.1 is a major update, with several new features and many bugfixes. Upgrading from WIEN2k_19 (or earlier) is highly recommended. WIEN2k-Update information: Current version: WIEN2k_21 Version 21 is a major update. It contains all bug fixes and patches as listed at the Github page of Gavin Abo and several others. For details see below. There are several new features as compared to version 19.2 and update is highly recommended. We want in particular thank Gavin Abo for various bug fixes, Pavel Ondracka for his efforts with OPENMP parallelization and fixing incompatibilities (gfortran) and Oleg Rubel for updating all Python scripts to version 3. Below are the most important changes compared to WIEN2k_19.2: We have recently published a review paper on WIEN2k and the official citation of WIEN2k should include: WIEN2k: An APW+lo program for calculating the properties of solids. P. Blaha, K.Schwarz, F. Tran, R. Laskowski, G.K.H. Madsen and L.D. Marks, J. Chem. Phys. 152, 074101 (2020) and WIEN2k, An Augmented Plane Wave + Local Orbitals Program for Calculating Crystal Properties (Karlheinz Schwarz, Techn. Universität Wien, Austria), 2018. ISBN 3-9501031-1-2 P. Blaha, K. Schwarz, G. K. H. Madsen, D. Kvasnicka, J. Luitz, R. Laskowski, F. Tran and L. D. Marks. OpenMP parallelization: This parallelization has been further improved (particular thanks to Pavel Ondracka ) and is very useful when running WIEN2k on modern multicore PCs. The programs dstart, 3ddens, nlvdw, optic (new), lapw0/1/2/so/ have been parallelized and run highly efficient on such cpus. Please note that a significant speedup of lapw1 is limited to 4 cores (due to limitations of the diagonalization using Intels mkl library). For more info please read the parallelization chapter of the Usersguide. local mBJ (lmBJ): enables the mBJ potential for multilayers, surfaces and molecules. It calculates a "local-c(r)" using a folding of grad rho/rho with a Gaussian of a certain width. kinetic energy tau in lapw2/lcore: direct tau calculation from grad |psi|*grad|psi| (not via vresp files in lapw0). More accurate numerically and the ONLY correct way when an additional non-local potential (like -orb or -eece) is used. mstar: new program (thanks to Oleg Rubel) calculating effective masses. See also: O.Rubel et al., Computer Physics Communications, 261 (2021), 107648. mixer: new version 10.8 by L.D.Marks, which should be again faster and more stable. It also features a new optimization under CONSTRAINTS for finding transition states or saddle point geometries. Please read the doku inside SRC_mixer or on our textbook page. optic: Massive speedup with OpenMP parallelization. rendos: new program calculates a "renormalized" DOS, where the interstital DOS is decomposed as much as possible and added to the partial DOS of the atoms. This allows a much better estimate of the DOS of localized (not changed) vs. delocalized states (like 3d vs. 4s PDOS of TM atoms). It can then be plotted using dosplot2 -ren (for an example see M.Bagheri and P. Blaha, J. Electron Spectroscopy and Related Phenomena, 230 (2019), 1 - 9. optimize_abc_lapw: a new script for finding the equillibrium lattice parameters of 2D (hexagonal or tetragonal) or 3D (orthorhombic) cases with relatively little effort. High parallelization possible. With an external loop (by hand) over an angle, even monoclinic lattices can be optimized now. At http://www.wien2k.at/reg_user/unsupported you can find a new tool by M. Jamal for 2D elastic constants (IRelast2D) all codes SRC_*/* have changed cmplx(a,b) to dcmplx(a,b), conjg to dconjg and in general gfortran (version 10) compatibility has been enforced. In addition, we have dropped fftpack and FFTW2 support and require now FFTW3 also for the sequential code. all csh-shell and Python scripts have been converted to /bin/tcsh and Python 3, respectively. check_minimal_software.sh: checks the minimal (and some recommended) software requirements for WIEN2k. It should be executed before an actual installation with siteconfing_lapw is done. The following updates have been made: VERSION_21: 10.4.2021 SRC: all scripts have been changed to /bin/tcsh analyse_lapw: support for :CONSTRAINT :VCOUL :FCG keywords Cgrace_dos_lapw: small fix for ymax setting check_minimal_software.sh: checks the minimal software requirements for WIEN2k. Should be executed before an actual installation with siteconfing_lapw is done. checkinwplot_lapw: Python 3 compatibility (mainly print function) create_elf_lapw: new script for ELF (electron localization function, default option), alpha or z plotting. Better accuracy than direct calculation of these quantities in lapw0. Needs tau files, executes internally lapw0, lapw5/3ddens and create_rho dosplot2: support for -pesb (broadened pes files) expand_lapw: Berrypi configuration changed, tcsh change with .change_tcsh hfpara_lapw: support for parallel hf-calculations with spin-orbit coupling instgen_lapw: support for ultra-heavy elements (Og) init_lapw: echos for default settings quoted, sanity check for -ecut -12 and 1.01; TETRA automatically replaced by TEMP for 2D systems; s/,/./ for german LANG setting init_orb_lapw: check for valid atom numbers for more atoms of the same type init_so_lapw: bug fix for multiple init_so calls with reducing symmetry because case.in2 was not updated. When multiplicity has changed, case.tausum instead of case.vresp is updated. joinvec_lapw: support for -hf -so and $scratch lapw0para_lapw: lapw0.error file created lapw1para_lapw: granularity set to 1 if $SCRATCH is something else than "./"; substitutions changed to ".output..."; sumw=0 checked (wrong .machines file) lapw2para_lapw: substitutions changed to ".output..."; support for case.tauval_NN lapwdmpara_lapw: lapwdm.error created, substitutions improved lapwsopara_lapw:substitutions improved, -norun (def-file only) support makestruct_lapw: small fix: title in quotes nlvdwpara_lapw: nlvdw.error file created nmr_analyse_lapw: uses summary_nmr_orb (from nmr_orb_analyse) and adds spin and dipolar terms for NMR in metals (for usage see UG) nmr_orb_analyse_lapw: new script for easier check of chemical shifts for several atoms, see UG (creates summary_nmr_orb) opticcopy_lapw: fix for spin-polarization opticpara_lapw: fix for empty SCRATCH variable, runs opticcopy for spin-polarization, substitutions improved optimize_abc_lapw: new script (together with xyzchange_lapw) allows for an efficient optimization of 2D (a,c in hex or tetragonal cases) or 3D (orthorhombic) lattice parameters (see UG). parabolfit_lapw: generalization for scf-file selection prepare_xsf_lapw: Python 3 compatibility (mainly print function) qtlpara_lapw: fix for parallel vectors on $SCRATCH reduce_rmt_lapw: bug fix for -vxc switch restore_lapw: support for tau files run*_lapw: set vresp variable at the beginning of the script, tau support, echo used instead of printf, .lcore usage printed into scf file, support for .forcedmat, .forceorb (uses unmixed dmat/vorb files) runeece_lapw: support for -fsm, -tau, -afm runfsm_lapw: support for -eece run_lapw, runsp_lapw: -dftd4, -lmbj support run_kgenhf_lapw: -newklist support save_lapw: -nodel (does not delete scf and broyden files for intermediate saves); tau* files saved, using cp -p (preserves date and time) scfmonitor_lapw: terminal qt support setrmt_lapw: RMT factor for P increased, nndist fix 3.3 instead of 2.3 siteconfig_lapw: support for FP_OPTS and OMP_SWITCH with corresponding Makefile.orig changes; -new option to come again to a "first installation"; FFTW3 mandatory; WIEN2k_parallel_options file; various small changes and improvements userconfig_lapw: BERRYPI variables removed vec2old_lapw: unaliases cp command wplot2xsf_lapw: Python 3 compatibility (mainly print function) write_inwf_lapw: Python 3 compatibility (mainly print function) x_lapw: various open of lapw1,2,sumpara,lapwso,qtl with irecl=-1 indication action='read'. aim: -dn broadening: -pes support create_rho support dftd4 support joinvec: support for -hf -so lapw0: -lmbj, -vsp xxx lapw2: unit 15, case.tmp with status=scratch removed; -tau lapw3: -up/dn -val/tot lapw5: -tau -exchange2 -sub -add lapw7: -so lapwso -p -d -hf lcore: -vresp -tau mstar, mstarqtl support mixer: support for -dftd4, -tau nlvdv: support for -lmbj optic: omp support qtl: case.tmpup/dn in qtl.def on scratch rendos support sumpara_tau support xyz2struct: support 3ddens: -tau -sub -add x_nmr_lapw: bug-fix for spin-orbit coupling: writing Emin (-10.) for case.inso corrected, support for -quota -so options x_nmr_quota_lapw: support for -quota -so SRC_3ddens: write_xsf.f: correction in writing xsf file. order: y,x,z --> x,y,z; bug fix for R-lattice in conventional (hexagonal) cell. SRC_aim: change from interst.frc to *.F, support for integration of a different array then the total charge density (spin densities for magnetic moments); XERROR --> XERROR1; speedup using vector-cos SRC_BerryPI: compatibility with python3, updated to newest version from github (WloopPHI for Weyl points; polarization calculations are now possible for any lattice type (previously limited to orthogonal lattice vectors); program options are changed to WIEN2k style) SRC_broadening: support for pes-broadening SRC_cif2struct: scan_cif.f: check for rounded 1/3,... for hex and rhomb SG SRC_clmaddsub: Bugfix in case that the K-vector order has changed SRC_dipan: dipan.f: bugfix for typo "lattice" in dirortho call SRC_dstart: Fix for "K-vectors up to GMIN" SRC_eosfit6: Better output, lists also error estimates of a,b,c SRC_Globals: modified W2kinit.F, W2kutils.c, dergl.f charge.f SRC_hf: Fix for "spin-polarized calc. + SO and no inversion symmetry (never really "active"  ,atoms up to ZZ=118SRC_joint: joint.f (symmat1 changed to symma1 for xmcd) SRC_kgen: basdiv.f (fix for manual setting of kx,ky,kz for CXZ lattice) SRC_lapw0: Support for staggerd magnetic fields (approximate "FSM" for antiferromagnets); kinetic energy density via tausum files (more accurate alternative to vrespsum), mandatory fftw3 usage (fftw2 and fftwpack removed), fftw*.f03 files removed, better omp-support, angular Lebedev integration (instead of Gauss-Legendre), accuracy of Coulomb potential increased, local-mBJ support, SCANL, R2SCAN, support for libxc-5.1.2, new mixing for mBJ, prepared for scf-MGGA (not active yet), atomnumber up to ZZ=118, fix for efg.f, vxclm2.f (fix for scaled PBE 1-d0->1.d0); rotdef.f SRC_lapw1: coors.f: improved NVECx check; removed print statements in dsyxev4.f, dsyrdt4.f, dsbein1.f; modules.F: fix for non-ELPA mpi version; open with -irecl=-1 SRC_lapw2: common radfu,potnlc,uhelp --> module radial_functions, uhelp; adaption for direct tau calculation: lapw2.F,l2main.F,atpar.F,outwin.f,fourir.F, ; fermi.f: check for zero denominator; lmax_to_dmft added to the qdmft module; open with -irecl=-1i; atom_mass till 118; fftw3 required SRC_lapwdm: diracout: ZZ till 118, fix for hf+so-vectors SRC_lapwso: open with -irecl=-1 (action='read'), diracout: Z-118; hmsec.f: matrix routine zheevr; kptout.f: k-vec with 1 more digit printed SRC_lcore: support for direct tau calculation SRC_lstart: modified case.inm; 2s-lo for small C spheres automatically removed from the basisset SRC_mixer: new version 10.8, constraints (saddle points, reaction barriers), see $WIENROOT/SRC_mixer/Docs; STIFF and STIFFER mode for difficult problems SRC_mstar: New program for effective masses in semiconductors, see O.Rubel et al., Computer Physics Communications 261 (2021) 107648. There is also a mstarqtl utility, which creates a "qtl" file with effective masses for "fat-band" spaghetti plots. SRC_nlvdw: support for local-mBJ (lmbj): calculates the local grad rho/rho folded with a Gaussian of width sigma for finite systems (surfaces, multilayers, molecules), mpi improvements, bug fix in calc_density: in cases with MULT>1 and pos and neag. IATNR, it may have used the wrong cubic/non-cubic charge summation SRC_nn: calculates also bond angles SRC_optic: *.frc --> *.F; speedup and omp parallelization (mmatrix.f, sph-UP.F,planew.F,lomain_op.F), fix for vecorhfup/dn, vectorhfso; opimat not written on unit 8 in symop.f SRC_optimize: filename corrections, comments in optimize.job updated SRC_pes: real*16 removed; changes for renormalized DOS (removing constrain for atoms with same name) SRC_qtl: array-reorganization, keeping qtl-array in memory, open with -irecl=-1 (action='read'), f-orbitals reassigned: fx(x2-3y2)-->fy(x2-3y2), fx(3x2-y2)-->fy(3x2-y2) because of different Ylm convention with (-1)**m SRC_rendos: new program for a renormalized DOS (without interstital) using a least squares fit such that the sum of the atomic PDOS gives the total DOS and the interstitial is removed. SRC_spaghetti: bz_lin.f: consistent use of toler (fixing additional black lines); inview.f: sizec_power (to scale sphere size of fat plots with some power), atom index = -1 in case.insp will read case.qtlmstar instead if case.qtl in spagh.f, circle scale plotted (which can be easily switch off by header=0 in case.insp or removed by editing the bounding box in the ps file) SRC_structeditor/SRC_readwrite: fix (multiplicity) for rwoctave.f SRC_sumpara: open with -irecl=-1 (action='read'), printing of diagonalized density matrix+eigenvectors SRC_symmetry: pstol introduced SRC_templates: changes in several input file templates: case.inaim (default phi and theta over all space), case.inb (PES), case.in5 (ADD), case.innlvdw (no default Kerneltype, sigma line), case.insp (header), case.int (KSEL=xx), .machines (omp_optic); new files: case.inm_tau, case.indftd4, case.incritic2 SRC_tetra: KSEL=XX in case.int limits integration to tetrahedra containing the selected k-point (for analysis). SRC_trig: new program create_rho.f (needed by create_elf_lapw); findMINcboa.f (fix for orthorhombic); fmax.f (fix for autorange of dosplot2); join_vectorfiles.f (-so -hf support); write_win_backend.f (format fix); xyz2struct.f (support for reading POSCAR files); SRC_w2w: modules_rc.F (allocation of E and projection); modw2w.f (mass up to 118) SRC_w2web: bin/w2web: fix for accept/deny keywords in w2web.conf (thanks to M. Kroecker, Univ. Freiburg); navig.pl (init_nlvdw); scf.pl, scfrun.pl (nl-vdW switch); min.pl (preference for -min over min_lapw) |
回复此楼» 猜你喜欢
统计一下：硕士毕业答辩后的谢师宴是学生出钱，还是老师出钱？
已经有20人回复
-
Scientific reports 投稿
已经有4人回复
-
银浆叠压烧结粘片的问题有没有好的解决方法
已经有5人回复
-
博士高校求职 安建大vs西科大
已经有13人回复
-
2024杰青和万人领军什么时候会评
已经有5人回复
-
希望嫡长子顺利出场！
已经有6人回复
-
职能部门工作人员态度不好是普遍的吗？怎么让他们态度好一些？
已经有5人回复
-
入职高校3年发表10+SCI，尽人事听天命
已经有48人回复
-
哎
已经有5人回复
-
选择
已经有11人回复
10