Copyright (C) 2009-2024 ABINIT group (XG)
This file is distributed under the terms of the GNU General Public License, see
~abinit/COPYING or
http://www.gnu.org/copyleft/gpl.txt .
For the initials of contributors, see ~abinit/doc/developers/contributors.txt .
Many thanks to the contributors to the ABINIT project
between January 2009 and May 2009. These release notes
are relative to improvements of ABINITv5.8 with respect to v5.7.
WARNING : This file is still rudimentary. Style and description
of contributions should be improved...
The list of contributors includes :
B. Amadon, P.M. Anglade, J.-M. Beuken, M. Boulanger, F. Bruneval,
D. Caliste, T. Deutsch, G.A. Franco, L. Genovese, M. Giantomassi,
F. Jollet, A. Lherbier, D. Hamann, M. Mancini,
S. Pesant, Y. Pouillon, T. Rangel, G.-M. Rignanese, M. Verstraete,
M. Torrent, J. Zwanziger
It is worth to read carefully
all the modifications that are mentioned in the present file,
and examine the links to help files or test cases ...
This might take some time ...
Please note the WARNINGS !
Xavier
__________________________________________________________________
Version 5.8, released on May 7, 2009.
Changes with respect to version 5.7 :
A. WARNINGS.
B. Most noticeable achievements (for users)
C. Changes in the package, for developers
D. Other changes (or on-going developments, not finalized)
__________________________________________________________________
A. WARNINGS
A.1 The default value for kptopt has been changed from 0 to 1 .
So, if you are interested in doing Gamma only calculations
(nkpt 1 kpt 0 0 0) you have to insert
kptopt 0
in your input file. On the contrary,
kptopt 1
is not needed anymore when calculations are done on the basis of
a grid of k point (with ngkpt or kptrlatt).
A.2 The default value for kptrlen has been changed from 20 to 30.
With the above-mentioned kptopt default change, the calculation
for a solid will proceed by default with a reasonable k point grid.
A.3 The default value for awtr has been changed from 0 to 1.
This only applies to GW calculations, when the screening is generated (optdriver=3).
A.4 In order to improve the portability of the reference tests,
the procedure to select the q points connecting two k points,
in the GW part of ABINIT, has been slightly modified. Now,
the coordinates of the q wavevectors that might be either 0.5 or -0.5
are preferably chosen to be 0.5, while this was undefined previously
(and in several cases, -0.5 was obtained). This induced a large modification
in all the GW reference files, although usually not at the level of the
quasiparticule energies. This is not always true, and modifications
as large as 0.01 eV were seen.
__________________________________________________________________
B. Most noticeable achievements
B.0 Note : several core developers have been working on the preparation of ABINITv6.0 ,
hence the noticeable achievements of the present v5.8. appear to be quite light.
B.1 The kinetic energy density is available, as well as the electron localisation function.
Availability for Norm-conserving pseudopotentials only, for
non-spin-polarized and spin-collinear cases (but not yet non-spin-collinear
case).
Input variables : prtkden, prtelf .
Test cases : v1#21 and v5#31 (for prtken - no test available for prtelf in v5.8.3)
Other documentation in doc/theory is available, see the ELF and KDEN subdirectories.
(Contribution from A. Lherbier)
B.2 In the PAW case, the band structure with band linewidth proportional to the projection of the
wavefunction on atomic partial waves (either L-decomposed or L,M-decomposed)
can be generated (fat-band band structure). See the input variable pawfatbnd.
Test v5#20.
(Contribution by B. Amadon)
B.3 Automatic calculation of the processor distribution for a parallel calculation:
determination of npkpt, npband, npfft and bandpp, by F. Jollet.
Documentation, automatic tests to be provided... (not easy for the last one).
The work on the band-FFT parallelism continues, a new preconditioner is available (F. Bottin).
Scalapack and MPI-IO improvements.
__________________________________________________________________
C. Changes for the developers
C.1 A script to check whether the input variables are documented (in doc/input_variables/*html),
of tested (in tests/*/Input/*in) has been written by Joe Zwanziger,
and can be activated by issuing
make abirules
in the tests directory.
C.2 There has been a complete renaming of the subdirectories of the src directory.
New adopted labelling scheme : xy_name , where xy is a two-digit integer number,
with the meaning of a level.
The directories are strictly ordered, the higher level ones calling
the lower level ones, excluding subdirectories with the same level.
Separate subdirectories have been created to isolated the routines
related to hiding the "write" operation, the "stop" operation, and the "timing"
operation. The definition of the global datastructures dtset and mpi_enreg
is now done at level 50 (50_abitypes).
C.3 The systematic constitution of modules containing the definition of objects and associated
methods has been started, by M. Giantomassi .
See the directory 27_toolbox_oop, as well as several m_* files in src/*
C.4 A new testing directory tests/gwdp has been defined. It holds reference
files obtained with GW calculations with double precision arithmetics (enable_gw_dpc).
__________________________________________________________________
D. Other changes
(or on-going developments, not yet finalized).
D.1 The computation of Raman scattering intensities, in the spin-polarized LDA case,
is under development. Work by Fabiana Da Pieve.
D.2 The use of Density Functional Perturbation Theory with PAW
(phonons, ddk, electric field, and magnetic field) is in preparation,
by M. Torrent, F. Jollet and J. Zwanziger.
D.3 Tests for the temperature dependence of the electronic structure (including the gap).
have been provided by P. Boulanger. See tests v5 #26,27,28 .
D.4 Self-consistent phonon calculations calculations have been implemented, by M. Verstraete.
Implemented is working, but portable tests and documentation should be provided.
Test (not yet portable) v5 #90
Input variables : scphon_temp and scphon_supercell
D.5 Continued work on GW by Matteo Giantomassi :
improvement of the treatment of the GW singularity, treatment of G0W0 on top of LDA+U,
treatment of the optical limit, Arnaud-Alouani treatment of exact expansion for
oscillators in sigma.F90
New input variables, not yet documented : gw_nqlwl, gw_qlwl , nomegasi, omegasimax.
No test either.
Additional improvement on GW (memory saving) by F Bruneval.
D.6 New test : v5#71, for one-shot GW with contour deformation technique.
(prepared by M. Giantomassi).
D.7 New test : v5#43, for very large grid in x, from PM Anglade.
D.8 New example configuration files, for MacOSX , Intel XEON, IBM, from JM Beuken.
D.9 New version of the BigDFT library (1.2.0.1) , from D. Caliste.
Work by D. Caliste, L. Genovese, T. Deutsch.
D.10 A new version of the build system, in which MPI is treated like the other libraries,
is in preparation, by Y. Pouillon
D.11 Work on the Python bindings, by D. Caliste
D.12 Improvements of the use of the Wannier90 library and associated doc (T. Rangel, M. Giantomassi)
In particular, now other file names than wannier90.win can be used.
D.13 Improvement of the interfacing with Wan-T going on. Use of Wan-T
in the many-body perturbation case (T. Rangel and G.-M. Rignanese).
D.14 M. Verstraete introduced the new input variable tsymgkq for anaddb. Test+ documentation to be provided.
D.15 New input variable for anaddb : tprt_yambo. Test + documentation + author to be provided.
D.16 Occupation matrix can be imposed within PBE0 (on-site), from FJollet.
D.17 Work on the recursion approach to high-temperature DFT calculations, by M. Mancini.
D.18 Numerous miscellaneous bug fixes (to the sources, as well as to the build system, including the plug-ins),
and improvements of documentation by B. Amadon, P.M. Anglade, J.-M. Beuken,
M. Boulanger, F. Bruneval, D. Caliste, G.A. Franco, M. Giantomassi, F. Jollet, D. Hamann, M. Mancini,
S. Pesant, Y. Pouillon, T. Rangel, M. Verstraete, M. Torrent, J. Zwanziger