Copyright (C) 2010-2013 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
March 2013 and August 2013. These release notes
are relative to improvements of ABINITv7.4 with respect to v7.2.
The list of contributors includes :
F. Abreu, B. Amadon, G. Antonius, J.-M. Beuken, F. Bottin, E. Bousquet, F. Bruneval, D. Caliste,
C. Espejo, L. Genovese, M. Giantomassi, Y. Gillet, X. Gonze, I. Lukacevic, M. Mancini,
A. Martin, C. Martins, S. Ponce, Y. Pouillon, A. Romero, M. Torrent, Bin Xu, 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 ...
Xavier
__________________________________________________________________
Version 7.4, released on August 10, 2013.
Changes with respect to version 7.2 :
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 AND IMPORTANT REMARKS
A.1 The new methodology for the test suite has been activated, and the old one has been disabled..
In practice, this means that a command like "make tests_acc" issued in the ~abinit/tests directory will not work anymore.
One should now use the script runtests.py . See runtests.py --help for the documentation about this new script.
See B.1 .
A.2 After the v7.4.1 was frozen, and released, two bugs in optics have been identified, and fixed.
They are related to the case when kptopt/=3 has been used in the ABINIT run before calling optics.
On one side, there was a misuse of the symmetries in case rprimd was not symmetric (the transpose of it should
have been used), which was e.g. very noticeable in case of the hexagonal crystalline system (but not, e.g. FCC).
On the other side, there was, if symmetry were used to decrease the number of k points, wrongly computed
weights for k points. The relative impact of the latter error was decreasing with the number of k point, so that
the converged calculation was correct.
Bug identified and fixed by Y. Gillet
__________________________________________________________________
B. Most noticeable achievements
B.1 The scripts for the automatic testing of ABINIT have been completely rewritten, with largely improved capabilities.
In particular, a subset of tests can be readily selected and executed, by using keywords.
Also, the number of threads can be varied, in order to obtain smaller wall clock time to full completion of the test suite.
The documentation is present in tests/doc and
http://www.abinit.org/developers/dev-environment/buildbot/howto-add-a-new-test-in-the-test-suite
The main script for testing is now called "runtests.py".
Get help with "runtests.py --help" , that provides a description of options, and typical examples.
The documentation has been updated.
Work by M. Giantomassi
B.2 The ABINIT GUI can be used in production.
The whole GUI is actually delivered now outside of the ABINIT package (it was too heavy).
The AbinitGUI binaries (but also sources) of the last stable version as well as additional material can be found at:
http://flavio.be/AbinitGUI.php
There you can download the pseudo-potential filetree organized for AbinitGUI (PSP.zip)
and other external post-processing libraries adapted for AbinitGUI (abipy.tar.gz & pymatgen.tar.gz)
Work by F. Abreu and Y. Gillet.
B.3 The APPA post-processor, dedicated to the analysis of Molecular Dynamics output files (trajectories),
has been made available.
Output files from large scale molecular dynamics simulations are not easy to manage.
All the data can be extracted from output files in NetCDF or ACSCII(text) formats.
APPA is a graphical tool written in python (PyQt4 library) able to read ABINIT output,
easy to use in order to follow a trajectory from molecular dynamics simulation
(total energy, pressure...).
It is also possible to calculate velocity autocorrelation function from the simulation,
radial pair distribution, vibrational density of states, etc....
See script/post-processing/appa, in particuler the README file
Work by A. Martin and F. Bottin (helped by M. Torrent)
B.4 The computation of orbital magnetism (also the Chern number) is now completed,
for use with PAW, with or without spin-orbit interaction.
Testing and debugging is underway.
By J. Zwanziger
B.5 icoulomb > 1 can be used with plane waves and PAW.
This has been tested.
This works in parallel (but only on FFTs).
Work in progress to use it within the KGB parallelization.
See test bigdft#23.
Work by T. Rangel + M Torrent, with the help of Luigi Genovese, Damien Caliste, and J.-M. Beuken
__________________________________________________________________
C. Changes for the developers
C.1 The abipy and gui projects have been transferred to GitHUB .
Please, see doc/gui/README and doc/abipy/README.
For abipy, see https://github.com/gmatteo/abipy .
For the AbinitGUI, see https://github.com/abinitgui/abinitgui.git
Work by F. Abreu, Y. Gillet, and M. Giantomassi
C.2 The new test scripts, see B.1, are now used for the test farm.
This means that the metadata (tolerances, launch commands, and documentation)
for each test are now present in the input file, and no more in the README, report.in or *.cnf files.
Most of the report.in and *.cnf files have been suppressed (not those in tests/buildsys and tests/abirules,
that still rely on old perl scripts.)
See the documentation at http://www.abinit.org/developers/dev-environment/buildbot/howto-add-a-new-test-in-the-test-suite
Work by M. Giantomassi, J.-M. Beuken, Y. Pouillon, and X. Gonze
C.3 Linked to B.1 and C.2, there has been a complete renaming of the tests
in the tests/mpiio, tests/paral, tests/seq directories, in order for the usual naming convention to be followed
with possibilities to address them one by one, or as a subset.
Other tests have been renamed or modified as well for this purpose, e.g. tdfpt_03, elphon (became eph),
Also, many obsolete keywords of the "make" command in ~abinit/tests have been disabled (e.g. make tests_acc has been disabled).
The documentatino has been updated accordingly.
Work by X. Gonze
C.4 The content of the util(/maintainers,/developers...) directory has been transferred
to the tests/Scripts or developers(/various or /maintainers) directory (or a few other relevant directories)
Work by Y. Pouillon
C.5 Created a new build subsystem for the doc .
Work by Y. Pouillon
C.6 New bots were added : coba2_intel13, green_gcc48, yquem_gcc47, yquem_gcc48, yquem_gcc47_mpich3, and yquem_g95
Work by J.-M. Beuken
__________________________________________________________________
D. Other changes
(or on-going developments, not yet finalized).
D.1 New version of libraries are now interfaced with ABINIT :
libxc-2.0.2
bigdft-abi-1.0.4
atompaw-3.1.0.2
Work by J.-M. Beuken
D.2 New unitary tests have been developed for FFT libraries :
- for the FFTW3 library (seq and threaded version)
- for the MKL library
- for the SG library
See the tests in the tests/unitary directory : tfftfftw3, tfftmkl, tfftsg.
Work by M. Giantomassi
D.3 The zero-point motion and temperature dependence of the gap (and electronic structure)
has been validated by cross comparison with PWSCF+YAMBO (after removing a bug in the latter,
both ABINIT and PWSCF+YAMBO agree for the temperature dependance of the gap of diamond, within a few meV).
The dynamical corrections to these effects has now been coded.
The postprocessing utility temperature.py is to be used for this purpose.
See Tv7#55 and 56.
Work by S. Ponce.
D.4 PAW + Bigdft project (beta version).
The new input variable wvl_ngauss has been created.
New Tests bigdft#32 and bigdft_paral#tbig_1 to test has been created to test wvl_ngauss.
A few patches have been added to the bigdft lib (bigdft-abi-1.0.5) to correct PAW + wavelets routines.
The creation of the ABINIT PAW library (libpaw) is going on, see src/42_libpaw .
Work by T. Rangel + M Torrent, with the help of Luigi Genovese, Damien Caliste, and J.-M. Beuken
D.5 The vdW-DF1 and DF2 are in debugging phase.
New input variables have been introduced :
vdw_df_damin, vdw_df_phisoft, vdw_df_ngpts,
vdw_df_rsoft, vdw_df_tolerance, vdw_df_tweaks .
There is also a new test to insure abirules compliance of VdW-DF input variables, see tests/vdwxc#03 .
From Y. Pouillon and C. Espejo
D.6 Work on BSE part of ABINIT, trying to improve the scaling, by Y. Gillet.
D.7 A new tutorial, describing the computation of properties at the nuclear site (EFG, NMR, NQR),
has been initialized. At present, the EFG part is available.
See doc/tutorials/lesson_nuc.html .
See also the new test tutorial#tnuc_1
Work by J. Zwanziger.
D.8 The application of magnetisation constraints (the magnetisation within a sphere must equal a predefined value)
is progressing. New input variables : magconon and magcon_lambda .
New test : v7#05 .
Work by M. Verstraete, I. Lukacevic and E. Bousquet.
D.9 Work on the implementation of the LOTF "Learn-On-The-Fly" technique.
The scheme is based on the idea of augmenting a unique,
simple parametrized force model by incorporating in it, at run time,
the quantum-mechanical information necessary to ensure accurate trajectories.
At present, only the glue model for gold is available.
Input variables : lotf_classic, lotf_nitex, lotf_nneigx, lotf_version.
Parallelism is working.
See test/paral#41 .
Work by M. Mancini
D.10 A new test for parallelism with BSE has been created, tests/paral#76 .
By Matteo Giantomassi
D.11 Comment lines are now unchanged by abiauty.pl
By X. Gonze .
D.12 Continued efforts on the development of Hartree-Fock
Hartree-Fock calculations are now available within Norm-Conserving PSP formalism
with the use of symetries and parallelism (beta version)
PAW + Hartree-Fock under development
From C. Martins (helped by M. Torrent and B. Amadon)
D.13 The Silvestrelli VdW approach has been further developed, with the coding of the
new flavor VdW-WF2 (A. Ambrosetti and P. L. Silvestrelli, doi:10.1103/PhysRevB.85.073101).
Amalgamation of close MLWFs and splitting of p-like Wannier functions (L. Andrinopoulos, N. D. M. Hine,
and A. A. Mostofi, doi:10.1063/1.3647912 JChemPhys.135.154105) have been enabled.
See the input variable vdw_xc=11
From C. Espejo and A. Romero
D.14 For the genetic algorithm, the coordinates can now be fixed.
Work by A. Romero
D.15 Make usexcnhat=0 work with non-collinear magnetism
Work by M. Torrent
D.16 icutcoul=6 (auxiliary function integration for 3D systems from P. Carrier et al., PRB 75, 205126 (2007))
is now the default for GW calculations.
Work by X. Gonze
D.17 The Krack table of norm-conserving HGH pseudopotentials for GGA has been added in psps/tables/NC_GGA_HGHK .
By M. Giantomassi
D.18 J. Zwanziger has provided a document about how to get a Crystal Structure into Abinit.
See doc/users/xtalinput.pdf .
D.19 Working on e-p Doppler broadening, by M. Torrent, J. Wiktor and G. Jomard.
D.20 Patch from Shauna Story in Rehr group, for output to file of the IFC and real space atomic positions
Incorporated by M. Verstraete
D.21 Continued efforts on the development of the abipy library
From M. Giantomassi and G. Antonius
D.22 Numerous miscellaneous additional bug fixes (to the sources, as well as to the build system,
including patches for the fallbacks), and improvements of documentation by :
F. Abreu, G. Antonius, J.-M. Beuken, E. Bousquet, F. Bruneval, C. Espejo,
M. Giantomassi, Y. Gillet, X. Gonze, Y. Pouillon, M. Torrent, Bin Xu, J. Zwanziger