EXCITING is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
EXCITING is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with EXCITING; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
EXCITING is still in its beta state so check the consistency of your results
carefully! Features marked as
are for
testing purposes only and should NOT be used for production work. The latest
releases can be
exciting-0.9.224
-After lots of careful testing, Anton Kozhevnikov found a serious bug in the
rotation of the first-variational eigenvectors. This has now been fixed - it
affects only calculations which require the non-reduced k-point set and only
then for certain symmetries, these include Hartree-Fock, OEP and optics.
-Shigeru Suehara implemented Pulay mixing (use mixtype=2), this greatly
improves the convergence rate of many calculations. Please be careful when
using this with magnetic calculations.
-F. Cricchio, F. Bultmark and L. Nordström improved the 3D plotting by
allowing for arbitrary box-like regions in both real and reciprocal space
(similar to plot2d)
-FC, FB and LN also improved the Fermi surface plotting in the same way, this
now also resolves the Fermi surface into spin-up and down for the collinear
magnetic case
-GGA now works with spin-orbit coupling, but it should not be used with
general non-collinear magnetism
-species files regenerated with slightly lower default semi-core local-orbital
linearisation energies
-removed unnecessary local-orbitals from species files
-iterative solver greatly improved, now also requires a step-length:
tauseq
-second-variational eigenvalues are no longer sorted for the collinear
magnetic case
-new examples added
-minor bug fixes and optimisations
Notes for developers
-the arguments to genppts have changed to allow for generation in arbitrary
boxes
exciting-0.9.219-223
-internal releases only
exciting-0.9.218
-fixed problem with forces for spin-polarised systems
-forces now work with LDA+U (
experimental)
-electron-phonon coupling now available, see the
Nb-superconductor example
(
experimental)
-calculation of phonon linewidths now available
(
experimental)
-calculation of superconducting T
C using the McMillan-Allen-Dynes
equation now available (
experimental)
-changed phonon calculation
DYN_... files: each computer now
calculates a separate row in the dynamical matrix
-phonon calculations slightly faster (you also need to run a ground state
calculation before the phonon run)
-thermodynamic properties calculated from phonon DOS now available: free energy,
entropy, heat capacity
-included phonon example for silicon (with results)
-S. Sharma updated and extended RDMFT
-calculation of
⟨
i,
k+
G+
q
| exp[
i(
G+
q)⋅
r] |
j,
k ⟩
matrix elements now available, use
task=130
(
experimental)
-energy loss near edge structure (ELNES) now available, see
BN-ELNES
example
-changed linear optics calculation from an integration over the Brillouin
zone to a direct sum over k-points
-optical relaxation time is now determined by
1/swidth
-fixed problem with MOKE as well as the off-diagonal elements of the
dielectric tensor and conductance
-sped up and simplified the diagonalisation of the spin density matrix
(Kübler trick) thanks to Lars Nordström
-F. Cricchio fixed and improved the calculation of the expectation values of
L,
S and
J by using the the density matrix
-F. Cricchio, F. Bultmark and L. Nordström fixed minor problem with
LDA+U energy
-F. Cricchio added the von Barth and Hedin exchange-correlation functional:
xctype=5
-fixed problem of non-convergence of atomic code (thanks to Stephan
Sagmeister)
-conjugate gradient solver for the first-variational problem is now available:
use
tseqit=.true. (see example
H2O-iterative)
-conjugate gradient solver allows for efficient fine-grained parallism with
OpenMP (scales linearly with number of processors)
-spherical covering is now more accurate
-code now runs with zero atoms for studying the electron gas (see examples
HEG-HF and
HEG-OEP for the Hartree-Fock and optimised
effective potential homogeneous electron gas)
-Poisson solver is now faster
-nucleus can now have finite size (set
ptnucl=.false.)
-energy printed out to more significant figures
-added MgB
2 Fermi surface example:
MgB2-Fermi-surface
-added optimised effective potential (OEP) example:
Si-OEP
-added 3D density plot example:
diamond-rho3D
-added example for plotting the
eg and
t2g contributions to the partial DOS:
TiC-eg-t2g
-changed variable
rlambda to
radkpt
-improved adaptive mixing scheme, also removed variable
betamax
-units of density of states is now states/Hartree/unit cell
-replaced cpu_time with system_clock calls
-lots of optimisations and simplifications
Notes for developers
-IMPORTANT: the first-variational spinor index and the k-point index have been
swapped in the following arrays: ngk, igkig, vgkl, vgkc, gkc, tpgkc, sfacgk
-an additional argument (tfbz) has been added to the routine genppts
exciting-0.9.152-217
-internal releases only
exciting-0.9.151
-fixed serious
bug
discovered by C. Koehler in the density symmetrisation routine - this
affected a limited number of spacegroups (like that in the B12 example),
and typically produced a warning about the total charge
-no other changes
exciting-0.9.150
-Reduced density matrix functional theory (RDMFT) now available thanks to
S. Sharma, see arXiv:0801.3787v1 [cond-mat.mtrl-sci]
(
experimental)
-Bug in symmetrisation of non-collinear magnetisation found and fixed by
F. Cricchio, F. Bultmark and L. Nordström
-F. Bultmark added new input variable
reducebf which reduces the
external magnetic field by this value after each iteration, thus at the
end of the s.c. loop the external magnetic fields are zero. By default
there is no reduction. See the manual for details.
-OEP calculations are now considerably faster
-Hartree-Fock now works properly with magnetism
-Alexey Baranov suggested an improvement of the cubic spline routine and
submitted an example on how to obtain perfectly smooth ELF plots (see
Ru-ELF)
-various minor changes to formatting of output files
-added more code documentation
-minor improvements and optimisations
exciting-0.9.146-149
-internal releases only
exciting-0.9.145
-fixed bug in non-collinear LDA+U, spotted by Francesco Cricchio and Lars
Nordström, and simplified code
-fixed problem which occurs with Intel Fortran version 10, thanks to
C. Koehler
-made the code thread safe for FFTW, thanks again to C. Koehler for pointing
this out
-minor optimisations
-new mailing list available for developers, administered by C. Meisenbichler
exciting-0.9.140-144
-internal releases only
exciting-0.9.139
-fixed bug which prevented FSM calculations converging (no other changes)
exciting-0.9.138
-LDA+U now available thanks to the considerable efforts of Francesco Cricchio,
Fredrik Bultmark and Lars Nordström, as well as the kind hospitality of
the group of Hardy Gross. There are three different methods implemented:
1. fully localised limit (FLL), 2. around mean field (AFM), and 3. an
interpolation between the two (see PRB 67, 153106 (2003) and PRB 52,
R5467 (1995)). The implementation is fully non-collinear and works with
spin-orbit coupling and spin-spirals. It is currently an
experimental feature. See the LDA+U examples and
documentation for details on how to use it.
-F. Cricchio added a fixed local moments feature, so that the moment
in every muffin-tin can be fixed. Note that because of this, the variable
fixspin has changed from Boolean to integer. See the manual for
details.
-planar spin-spirals are no longer experimental and may be used for production,
thanks to extensive testing by F. Cricchio (the conical kind still need more
checking). FC and L. Nordström also found and fixed a bug in the
spin-spiral symmetry setup.
-after extensive checking Fritz Körmann found a problem with the GGA
implementation which has now been fixed: may have a small effect on magnetic
GGA calculations
-a new GGA functional
PBEsol is
available which "restores the density-gradient expansion for exchange in
solids and surfaces": use
xctype=22
-F. Bultmark added a warning message for when metals run out of empty states
-Christian Meisenbichler has created a Git repository where you can compare
different versions of the code. See the
Wiki for
details.
-added compensating background charge energy to total when
chgexs≠0
-lots of improvements, optimisations and minor bug fixes
exciting-0.9.115-137
-internal releases only
exciting-0.9.114
-changed the Hartree-Fock routine (hopefully for the last time): use
task=5 to run it, still an
experimental
feature
-added Hartree-Fock example
-fixed bug in spin-orbit coupling: affects the splitting of low-lying states
-many minor improvements, optimisations and fixes
-speed up of APW-local-orbital kinetic energy surface term thanks to
Francesco Cricchio
-intraband Drude-like contribution to dielectric tensor now added to total when
intraband=.true.
-added optics example for aluminium: Al-linopt
-upgraded to LAPACK 3.1.1 (still works fine with 3.0)
-EXCITING
Wiki now up and
running
