1.. _changelog: 2 3Change Log 4========== 5 6Mar-17-2021: Version 1.22.3 7--------------------------- 8- Fix ``--read-gamma`` option to work. 9 10Feb-21-2021: Version 1.22.2 11--------------------------- 12- Fix PyPI source distribution package 13 14Feb-21-2021: Version 1.22.1 15--------------------------- 16- ``phono3py`` command didn't work. This was fixed. 17- Fix behaviour when specifying ``--thm`` and ``--sigma`` simultaneously. 18 19Jan-29-2021: Version 1.22.0 20--------------------------- 21- Maintenance release to follow phonopy v2.9.0. 22 23Sep-30-2020: Version 1.21.0 24--------------------------- 25 26- Maintenance release to follow the change of phonopy at v2.8.1 27- Improvements of phono3py loader (``phono3py.load``), ``phono3py-load`` 28 command, API, and ``phono3py_disp.yaml``. 29- Harmonic phonon calculation on mesh was multithreaded. This is 30 effective when using very dense mesh with non-analytical term 31 correction (probably rare case). 32- Real and imaginary parts of self energy and spectral function of 33 bubble diagram at API level 34 35Mar-3-2020: Version 1.20.0 36-------------------------- 37 38- ``phono3py_disp.yaml`` is made when creating displacements in 39 addition to ``disp_fc3.yaml`` and 40 ``disp_fc2.yaml``. ``phono3py_disp.yaml`` will be used instead of 41 ``disp_fc3.yaml`` and ``disp_fc2.yaml`` in the future major release 42 (v2.0). 43 44Mar-3-2020: Version 1.19.1 45-------------------------- 46 47- Release for pypi and conda (atztogo channel) packagings 48 49Mar-2-2020: Version 1.19.0 50-------------------------- 51 52- Improvements of phono3py loader and API. 53- Improvement of interfaces to calculators. Now it is expected to be 54 much easier to implement calculator interface if it exists in 55 phonopy. 56- Fixed dependency to phonopy v2.6.0. 57 58Dec-22-2019: Version 1.18.2 59--------------------------- 60 61- Initial version of phono3py loader (``phono3py.load``) was 62 implemented. See docstring of ``phono3py.load``. 63 64Oct-17-2019: Version 1.18.1 65--------------------------- 66 67- Fix of phono3py-kaccum to follow the latest phonopy. 68 69Oct-9-2019: Version 1.18.0 70--------------------------- 71 72- Maintenance release 73 74Apr-18-2019: Version 1.17.0 75--------------------------- 76- ``--cfz`` option was made to subtract residual forces. See 77 :ref:`cfz_option`. 78- ``--cutoff-pair`` was made to override the cutoff pair distance 79 written in ``disp_fc3.yaml`` when using on calculating force 80 constants. This is useful when checking cutoff distance 81 dependency. So the use case of having fully computed ``FORCES_FC3`` 82 is assumed. 83- TURBOMOLE interface is provided by Antti Karttunen 84 (``--turbomole``). 85- Compatibility of ``fc2.hdf5`` and ``force_constants.hdf5`` was 86 improved for all calculators to store physical unit information in 87 the hdf5 file. See :ref:`file_format_compatibility`. 88 89Mar-24-2019: Version 1.16.0 90--------------------------- 91- Bug fixes and catching up the updates of phonopy. 92- Most of hdf5 output files are compressed by ``gzip`` as 93 default. This compression can be set off, see 94 :ref:`hdf5_compression_option`. 95- (Experimental) ``phono3py`` command accepts ``phono3py.yaml`` type 96 file as an input crystal structure by ``-c`` option. When ``DIM`` 97 and any structure file are not given, ``phono3py_disp.yaml`` 98 (primary) or ``phono3py.yaml`` (secondary) is searched in the current 99 directory. Then ``phono3py.yaml`` type file is used as the input. 100 By this, semi-automatic phono3py mode is invocked, which acts as 101 102 (1) ``supercell_matrix`` corresponding to ``DIM`` in the 103 ``phono3py.yaml`` type file is used if it exists. 104 (2) ``phonon_supercell_matrix`` corresponding to ``DIM_FC2`` in the 105 ``phono3py.yaml`` type file is used if it exists. 106 (3) ``primitive_matrix`` in the ``phono3py.yaml`` type file 107 is used if it exists. Otherwise, set ``PRIMITIVE_AXES = AUTO`` 108 when ``PRIMITIVE_AXES`` is not given. 109 (4) NAC params are read (``NAC = .TRUE.``) if NAC params are 110 contained (primary) in the ``phono3py.yaml`` type file or if 111 ``BORN`` file exists in the current directory (secondary). 112 113Nov-22-2018: version 1.14.3 114---------------------------- 115- Update to work with phonopy v1.14.2. 116- Ph-ph interaction can be read (``--read-pp``) and write 117 (``--write-pp``) in RTA thermal conductivity calculation, too. Mind 118 that the data stored are different with and without 119 ``--full-pp``. Wihtout ``--full-pp`` the data are stored in 120 complicated way to save data side, so it is not considered readable 121 by usual users. 122 123June-20-2018: version 1.13.3 124---------------------------- 125 126- ``--lw`` (linewidth) option was removed. Use ``--br`` option and 127 find 2*gamma values as linewidths in ``kappa-xxx.hdf5`` file. 128- Documentation of ``--lbte`` option is available at 129 :ref:`direct_solution`. 130- This version is dependent on phonopy>=1.13.2. 131 132May-17-2018: version 1.13.1 133---------------------------- 134 135- Compact force constants are implemented (See :ref:`compact_fc_option`). 136 137Mar-16-2018: version 1.12.9 138---------------------------- 139 140- Definition of ``mode_kappa`` values in output hdf5 file is 141 changed. Previously they were divided by number of grid points, but 142 now not. Therefore users who compute ``kappa`` from ``mode_kappa`` 143 need to be careful about this change. This does not affect to 144 ``phono3py-kaccum`` results. 145 146Feb-1-2018: version 1.12.7 147---------------------------- 148 149- ``--tsym`` option is removed. Now with ``--sym-fc3r`` and 150 ``--sym-fc2`` options, 151 translational invariance symmetry is also applied. 152- ``--sym-fc`` option is added. This is just an alias to specify both 153 ``--sym-fc3r`` and ``--sym-fc2`` together. 154- Documentation on ``--write-phonon`` and ``--read-phonon`` options is 155 written. These options are used to save harmonic phonon infromation 156 on strage. 157 158Nov-22-2017: version 1.12.5 159----------------------------- 160 161- Bug fix of RTA thermal conductivity. This bug exists from version 162 1.10.11.18 (git e40cd059). This bug exhibits when all the following 163 conditions are met: 164 165 1. RTA thermal conductivity calculation. 166 2. Tetrahedron method for Brillouin zone integration is used. 167 3. Number of triplets is smaller than number of bands at each grid point. 168 4. Without using ``--full-pp``. 169 170 171 (3) happens when the primitive cell is relatively large. Number of 172 triplets can be shown using ``--stp`` option. A race condition of 173 OpenMP multithreading is the source of the bug. Therefore, if it 174 occurs, the same calculation comes up with the different thermal 175 conductivity value in every run time, for whcih it behaves like 176 randomly. 177 178- RTA thermal conductivity with smearing method (``--sigma``) is made 179 to run with smaller memory consumption as similar as tetrahedron 180 method (``--thm``). 181 182Nov-17-2017: version 1.12.3 183---------------------------- 184 185- Command option parser of the phonopy tools is replaced from 186 ``optparse`` to ``argparse``. 187- The filenames used with these options were the positional arguments 188 previously. Now they are the command-line arguments, i.e., filenames 189 have to be put just after the option name like ``-f vasprun.xml-001 190 vasprun.xml-002 ...``. 191- The names of auxiliary tools (``kdeplot`` and ``kaccum``) are 192 changed, for which the prefix phono3py- is added to the old names to 193 avoid accidental conflict with other script names already existing 194 under bin directory. 195- :ref:`sigma_cutoff_option` option was created. 196 197Jun-18-2017: version 1.11.13 198---------------------------- 199 200- :ref:`normal_umklapp_option` option was made. 201- Many minor updates: fixing bugs, improving usabilities. 202- Improve of :ref:`auxiliary_tools_kaccum` and :ref:`auxiliary_tools_kdeplot`. 203 204Mar-31-2017: version 1.11.11 205---------------------------- 206 207- Abinit code interface is implemented and now under the testing. 208- Reduction of memory usage in RTA thermal conductivity 209 calculation. This is especially effective for larger unit cell 210 case. Currently combinations with --full_pp, --write_gamma_detail, 211 and --simga(smearing method) are not supported for this. Performance 212 tuning is under going. In some case, computation can be slower than 213 the previous versions. 214 215Feb-9-2017: version 1.11.9 216--------------------------- 217 218- This version works coupled with phonopy-1.11.8 or later. 219- CRYSTAL code interface is implemented by Antti J. Karttunen. 220 221Dec-14-2016: version 1.11.7 222------------------------------ 223 224- This is a maintenance release. This version must be used with 225 phonopy-1.11.6 or later. 226 227Nov-27-2016: version 1.11.5 228------------------------------ 229 230- ``gaccum`` is merged to ``kaccum``. ``gaccum`` is removed. See 231 :ref:`auxiliary_tools_kaccum`. 232- ``kdeplot`` is added. See :ref:`auxiliary_tools_kdeplot`. 233 234Apr-24-2016: version 1.10.9 235------------------------------ 236 237- Failure of writing ``kappa-mxxx-gx.hdf5`` was fixed. 238 239Apr-16-2016: version 1.10.7 240------------------------------ 241 242- API example is prepared and it is found in ``Si`` example. No 243 doucment yet. 244- Si pwscf example was placed in ``example-phono3py`` directory. 245- User interface bug fix. 246 247Mar-15-2016: version 1.10.5 248------------------------------ 249 250- Numbering way of phono3py version was just changed (No big updates 251 were made against previous version.) The number is given based on 252 the phonopy version. For example, the harmonic part of 253 phono3py-1.10.5 is based on the code close to phonopy-1.10.4. 254- Python3 support 255- For the RTA thermal conductivity calculation mode with using the 256 linear tetrahedron method, only necessary part of phonon-phonon 257 interaction strengh among phonons. This improves lifetime 258 calculation performance, but as the drawback, averaged ph-ph 259 interaction strength can not be given. See :ref:`full_pp_option`. 260- Pwscf interface (:ref:`calculator_interfaces`) 261 262Oct-10-2015: version 0.9.14 263------------------------------ 264 265- Computational performance tuning for phonon-phonon interaction 266 strength calculation was made by Jonathan Skelton. Depending on 267 systems, but 10-20% performance improvement may be possible. 268- ``--stp`` option is created to show numbers of q-point triplets to 269 be calculated. See :ref:`command_options`. 270- ``--write_gamma`` and ``--read_gamma`` support using with ``--bi`` 271 option. Therefore a thermal conductivity calculation can be 272 distributed over band index, too. This may be useful for the system 273 whose unit cell is large. 274 275Sep-26-2015: version 0.9.13 276------------------------------ 277 278- Changed so that ``--wgp`` option writes ``grid_address-mxxx.hdf5`` 279 instead of ``grid_address-mxxx.dat``. 280- ``--write_detailed_gamma`` is implemented. See :ref:`command_options`. 281- When running without setting ``--thm`` and ``--sigma`` options, 282 linear tetrahedron method corresponding to ``--thm`` is used as the 283 default behavior. 284- ``--ise`` options is created. 285 286Aug-12-2015: version 0.9.12 287------------------------------ 288 289- Spglib update to version 1.8.2.1. 290- Improve computational performance of ``kaccum`` and ``gaccum``. 291 292Jun-18-2015: version 0.9.10.1 293------------------------------ 294 295- Bug fix of ``gcaccum`` 296 297Jun-17-2015: version 0.9.10 298---------------------------- 299 300- Fix bug in ``kaccum``. When using with ``--pa`` option, irreducible 301 q-points were incorrectly indexed. 302- ``gaccum`` is implemented. ``gaccum`` is very similar to ``kaccum``, 303 but for :math:`\Gamma_\lambda(\omega_\lambda)`. 304- spglib update. 305 306Changes in version 0.9.7 307------------------------- 308 309- The definition of MSPP is modified so as to be averaged ph-ph 310 interaction defined as :math:`P_{\mathbf{q}j}` in the arXiv 311 manuscript. The key in the kappa hdf5 file is changed from ``mspp`` 312 to ``ave_pp``. The physical unit of :math:`P_{\mathbf{q}j}` is set 313 to :math:`\text{eV}^2`. 314 315Changes in version 0.9.6 316------------------------ 317 318- Silicon example is put in ``example-phono3py`` directory. 319- Accumulated lattice thermal conductivity is calculated by ``kaccum`` 320 script. 321- JDOS output format was changed. 322 323Changes in version 0.9.5 324------------------------- 325 326- In ``kappa-xxx.hdf5`` file, ``heat_capacity`` format was changed 327 from ``(irreducible q-point, temperature, phonon band)`` to 328 ``(temperature, irreducible q-point, phonon band)``. For ``gamma``, 329 previous document was wrong in the array shape. It is 330 ``(temperature, irreducible q-point, phonon band)`` 331 332 333Changes in version 0.9.4 334------------------------ 335 336- The option of ``--cutoff_mfp`` is renamed to ``--boundary_mfp`` and 337 now it's on the document. 338- Detailed contribution of ``kappa`` at each **q**-point and phonon 339 mode is output to .hdf5 with the keyword ``mode_kappa``. 340 341Changes in version 0.8.11 342------------------------- 343 344- A new option of ``--cutoff_mfp`` for including effective boundary 345 mean free path. 346- The option name ``--cutfc3`` is changed to ``--cutoff_fc3``. 347- The option name ``--cutpair`` is changed to ``--cutoff_pair``. 348- A new option ``--ga`` is created. 349- Fix spectrum plot of joint dos and imaginary part of self energy 350 351Changes in version 0.8.10 352------------------------- 353 354- Different supercell size of fc2 from fc3 can be specified using 355 ``--dim_fc2`` option. 356- ``--isotope`` option is implemented. This is used instead of 357 ``--mass_variances`` option without specifying the values. Mass 358 variance parameters are read from database. 359 360Changes in version 0.8.2 361------------------------ 362 363- Phono3py python interface is rewritten and a lot of changes are 364 introduced. 365- ``FORCES_SECOND`` and ``FORCES_THIRD`` are no more used. Instead just 366 one file of ``FORCES_FC3`` is used. Now ``FORCES_FC3`` is generated 367 by ``--cf3`` option and the backward compatibility is simple: ``cat 368 FORCES_SECOND FORCES_THIRD > FORCES_FC3``. 369- ``--multiple_sigmas`` is removed. The same behavior is achieved by 370 ``--sigma``. 371 372Changes in version 0.8.0 373------------------------ 374 375- ``--q_direction`` didn't work. Fix it. 376- Implementation of tetrahedron method whcih is activated by 377 ``--thm``. 378- Grid addresses are written out by ``--wgp`` option. 379 380Changes in version 0.7.6 381------------------------ 382 383- Cut-off distance for fc3 is implemented. This is activated by 384 ``--cutfc3`` option. FC3 elements where any atomic pair has larger 385 distance than cut-off distance are set zero. 386- ``--cutpair`` works only when creating displacements. The cut-off 387 pair distance is written into ``disp_fc3.yaml`` and FC3 is created 388 from ``FORCES_THIRD`` with this information. Usually sets of pair 389 displacements are more redundant than that needed for creating fc3 390 if index permutation symmetry is considered. Therefore using index 391 permutation symmetry, some elements of fc3 can be recovered even if 392 some of supercell force calculations are missing. In paticular, all 393 pair distances among triplet atoms are larger than cutoff pair 394 distance, any fc3 elements are not recovered, i.e., the element will 395 be zero. 396 397Changes in version 0.7.2 398------------------------ 399 400- Default displacement distance is changed to 0.03. 401- Files names of displacement supercells now have 5 digits numbering, 402 ``POSCAR-xxxxx``. 403- Cutoff distance between pair displacements is implemented. This is 404 triggered by ``--cutpair`` option. This option works only for 405 calculating atomic forces in supercells with configurations of pairs 406 of displacements. 407 408Changes in version 0.7.1 409------------------------ 410 411- It is changed to sampling q-points in Brillouin zone. Previously 412 q-points are sampled in reciprocal primitive lattice. Usually this 413 change affects very little to the result. 414- q-points of phonon triplets are more carefully sampled when a 415 q-point is on Brillouin zone boundary. Usually this 416 change affects very little to the result. 417- Isotope effect to thermal conductivity is included. 418 419Changes in version 0.6.0 420------------------------ 421 422- ``disp.yaml`` is renamed to ``disp_fc3.yaml``. Old calculations with 423 ``disp.yaml`` can be used without any problem just by changing the 424 file name. 425- Group velocity is calculated from analytical derivative of dynamical 426 matrix. 427- Group velocities at degenerate phonon modes are better handled. 428 This improves the accuracy of group velocity and thus for thermal 429 conductivity. 430- Re-implementation of third-order force constants calculation from 431 supercell forces, which makes the calculation much faster 432- When any phonon of triplets can be on the Brillouin zone boundary, i.e., 433 when a mesh number is an even number, it is more carefully treated. 434