Release History¶
Release Notes¶
Release note of SALMON ver. 2:
Release note of SALMON ver. 0 and 1:
Details of Minor Changes¶
The following is the history of fixed bugs and changes in models/inputs/outputs after releasing v.2.0.0. (not complete list currently)
Fixed bugs¶
(Fixed in v.2.2.1)
- The OpenACC mode by newer versions of Nvidia HPC SDK was not supported.
- GNU Compiler Collection (GCC) was not supported.
- The sign of the polarizability in isolated systems (yn_periodic=n) was wrong.
- The definition of the excitation energy for the frozen Hamiltonian calculation (yn_fix_func=y) was wrong.
- File output of the atomic force for the multiscale mode (e.g. theory=multi_scale_maxwell_tddft and yn_out_rvf_rt=y) was wrong.
- File output of
sysname_rt.data
for very large systems was wrong. - An array allocation for the projection_option mode was wrong.
- An array allocation for spin-noncollinear isolated systems (yn_periodic=n and spin=noncollinear) was wrong.
(Fixed in v.2.2.0)
- For the incident pulse for isolated systems (
yn_periodic=’n’
), the circular polarization was not defined and the CEP for Acos2 envelope was wrong. - For spin-unpolarized systems (
spin=‘unpolarized’
), the initial value of the occupation rate was wrong when the electron number (nelec
) is odd. - When
unit_system=‘A_eV_fs’
is specified,temperature
was mistakenly defined in the atomic unit. - For isolated systems (
yn_periodic=’n’
), the multipole expansion for the boundary conditions of the Hartree potential (Poisson equation) was wrong. - For periodic systems (
yn_periodic=’y’
),al_vec[123]
for orthogonal cells yielded unintended error. - For spin-noncollinear systems (
spin=‘noncollinear’
), the band-decomposition of the spin magnetization in ***_mag.data files was wrong.
(Fixed in v.2.1.0)
- Non-local term of the transition moment printed by the option of "yn_out_tm=y" has been fixed (just printing issue).
- Parallelization for orbitals for calculation of transition moment by "yn_out_tm=y" has been suported
- Bug of the segmentation fault occurred by "yn_ffte=y" with parallelization for orbitals using isolated system has been fixed.
- Reading of CIF file format for symmetry option has been improved.
- Combination of non-uniform user-defined k-points and symmetry option has been supported.
(Fixed in v.2.0.2)
- Small noise on the total energy in TDDFT calculation (that is seen with weak pulse around e.g. I=1d9 W/cm2) has been removed.
- The printed absolute values of electron density in cube format has been fixed..
- Printing of the external field in TDDFT calculation of the isolated system has a bug in v.2.0.0 and v.2.0.1. It has been fixed in v.2.0.2.
- The file reading option of the external electric field in TDDFT calculation ("file_input1" in &emfield) has been fixed.
- Invalid occupation number printed in SYSNAME_ovlp.data file for projection option with non-uniform k-points has been fixed.
(Fixed in v.2.0.0)
- The imaginary part of wavefunction was not printed in cube format until v.1.2.1
(Fixed in v.?.?.?)
- Abnormal calculation that sometimes happens if zero value is included in the atomic coordinate in the input with "A_eV_fs" has been fiexd.
Changes of models/inputs/outputs¶
(v.2.2.1)
- New input keyword for the preconditioning of CG method for accelerating the DFT computation
yn_preconditioning
- For the spin-noncollinear mode (spin='noncollinear'), some input keywords and output files are changed and added.
sysname_rt_spin.data
: the output file for the spin magnetization and spin current density.- Change the input keyword:
out_magnetization_step
-->out_rt_spin_step
. - New input keywords:
yn_out_mag_decomposed_rt
,yn_out_spin_current_decomposed
,yn_out_spin_current_micro
.
- New input keyword to read a
.cube
file of the initial electron density for accelerating the DFT computationmethod_init_density='read_dns_cube'
(v.2.2.0)
- New theory options for SBE and Maxwell-SBE are added.
theory = 'sbe'
theory = 'maxwell_sbe'
- Input keywords for the Poisson equation solver are added.
method_poisson
yn_fftw
- New TDDFT analysis options are added.
yn_fix_func
projection_option=‘td’
threshold_projection
yn_out_intraband_current
yn_out_current_decomposed
,out_current_decomposed_step
yn_out_micro_je
,out_micro_je_step
(v.2.1.0)
- Input variables for Spin-orbit coupling are added
- "yn_spinorbit"
- "spin = noncollinear"
- "out_magnetization_step"
- New options for calculation of dielectric function and conductivity based on transition moments at the end of the GS calculation is added. The related input variables are
- "yn_out_gs_sgm_eps"
- "out_gs_sgm_eps_mu_nu"
- "out_gs_sgm_eps_width"
(v.2.0.2)
- The definition of the total energy of the periodic system printed in TDDFT calculation has changed: The electric field energy is included until v.2.0.1. It has not been included from v.2.0.2.
- The directory names generated by "method_wf_distributor=slice" option have changed from v.2.0.2.