Skip to content

Latest commit

 

History

History
216 lines (181 loc) · 13.7 KB

changelog.md

File metadata and controls

216 lines (181 loc) · 13.7 KB
Raw

v1.1.0

  • Update to compute results shown in: https://arxiv.org/abs/2303.11344 .
    • Added particle_type = 'mdm', 'edm' options for calculation of the absorption rate of DM coupling to electron MDM/EDM.
    • Added a new type of electronic state approximation, 'elec_state_atomic_type', which has two sub-types corresponding to bound states expanded in an STO basis, 'elec_state_atomic_STO_basis_type', and proper continuum, Coloumb wave function states which solve V = -Z/r, 'elec_state_atomic_continuum_type'. These are suitable for describing an 'atomic target', see the above paper for more details.
      • These states are defined by the property that they can be written as \psi(x) = R(r) Y(\theta, \phi), where Y are spherical harmonics. Future subtypes may be incorporated if they provide a 'compute_radial_wf' and 'compute_wf' procedure.
      • These are currently only used in absorption rate calculations.
    • Added a new method for computing transition matrix elements between atomic states, in the q -> 0 limit.
      • Currently, only transitions from atomic states to atomic states are supported, i.e., there is no way to calculation bloch state -> atomic state, or vice versa. Additionally, only initial STO -> final continuum is supported since the STO basis provides a nice analytic formula for the first and second derivatives. Initial atomic STO -> final atomic STO may be added with relative ease, whereas initial continuum -> continuum would be more difficult (although it is unclear what circumstance that would be important).
    • Added new self-energy, and corresponding absorption rate, calculations for the MDM/EDM models.
    • Added atomic target and crystal target EDM absorption examples.
    • Updated input file documentation.

v1.0.0

v1.0.0-beta

  • New Electronic Configuration file
    • Contains all the relevant information about the electronic configuration, no more need to separate ‘core’, ‘free’, and ‘valence and conduction’ states
  • Extended absorption and dielectric calculations
    • Include all states (e.g. core and free)
  • Separated electronic state and calculation dependence
    • Generalized input electronic states, initial/final states can now be specified in different bases, i.e., PW basis, STO basis, or single PW’s (previously ‘free’ states)
      • For example, allows for calculation of STO basis → STO basis transitions
    • Easily add different bases
    • Only need to specify how matrix elements are computed for a specific approximation of the wave functions, the ‘particle physics’ part of the code base doesn’t need to change
  • Automatically generated input documentation
  • Increased use of CMake and preprocessor directives
    • Version number automatically passed to program and documentation
  • Use hdf5_utils for even easier access to data inside an hdf5 file
  • Building new documentation with Sphinx
  • Improved parallelizability

    • Everything is parallelized over total number of transitions

    • All calculations use MPI_Reduce commands

  • Removed dependence on k_weight which only caused confusion (sum to 1 or 2?)
  • New input format - Configuration File
    • Major reasons

      • Dynamically sized 1D arrays

        • No need to specify number of elements in an input array

      • Enforced defaults

      • Easily add documentation

      • Split in to as many files as you want

    • Minor reasons
      • No need for ! at end of file
  • Added many new examples which all use the new input, and electronic configuration files.

v0.3.0

  • Updated calculation of some self energies to avoid numerically small numbers. For example, the dielectric computed with either ( \Pi_{1, 1} ) or ( \Pi_{v^i, v^j} ) is identical.
  • Added the calculation of ( \Pi_{\mathbf{v} \cdot \sigma, \mathbf{v} \cdot \sigma} ).
    • Used in generalizing the pseudoscalar absorption rate for spin dependent wave functions.
  • Added a timer type for more granular timing data.
  • Complete rework of examples so each example has a specific ID and readme.md description file.
    • Removed tests in favor of a script which runs all the examples, examples/run_all_examples.sh.
  • Speed improvement of the dielectric calculation in the scattering ( ( q \gg \omega ) ) regime.

v0.2.8

  • Minor updates to dm_model and PW_dataset types
  • Added routine to find the maximum momentum transfer within the 1BZ, get_q_max_1BZ.
  • Generalized dielectric inputs, anisotropic dielectrics in each (q, omega) bin are now accepted. See the Dielectric Data File documentation for more details.
  • Added cc_ext calculation mode, a new way to compute the core -> conduction contribution which allows the q range to extend to the kinematic bound.
  • Added a new integration routine when going from differential quantities such as ( \frac{dR}{d \omega} ), ( \frac{dR}{dq} ) to the binned values.
  • Added timing information to printed and saved output.
  • Updated documentation.

v0.2.7

  • Fixed bug with Ef_max cutoff by shifting valence band maximum to be the zero-point energy.
  • Updated build system to CMake, removing FoBiS entirely.
  • The OpenMPI subroutine, MPI_Reduce is used when appropriate verus manual send and recieve calls.
    • Currently implemented in scattering rate and dielectric calculations
  • Temporary file is deleted when creating the core electron configuration file with utilities/core_elec_config/create_elec_config.ipynb
  • Removed TODO.md which was not being used (may reintroduce in the future if it becomes useful).
  • Updated the binned_wfc utility.
  • Added documentation for the binned_wfc utility.
  • Optimized dielectric calculation, improved binned procedure to save memory.
  • Update to documentation and refactored main folder.
  • Added home page to README.md.

v0.2.6

  • Added the option to build with cmake. This will become the default build option in v1.0.0. Installation instructions can be found in install-cmake.md or on the website.
  • Communication of scattering rate data is now done with an MPI_Reduce command versus manual send and recv calls.
  • Added a timing subroutine when computing the dielectric.
  • Miscellaneous updates to documentation.
  • Minor bug fixes: some variables were not saving to the output files.

v0.2.5

  • Major code refactor and other miscellaneous improvements on the way to v1.0.0
    • Changed valence -> conduction parallelization scheme from {i, i'} -> {i, k}, allowing for significant speedups when ( number of valence bands ) x ( number of conduction bands ) < ( number of processors )
    • Added n_val_max, n_cond_max options to specify the maximum number of valence and conduction bands to keep in the calculation (valence -> conduction scattering, absorption, and dielectric calculations). n_val_max counts down from the Fermi surface and n_cond_max counts up from the Fermi surface.
    • Changed core -> conduction parallelization scheme from {i, i'} -> {i, kf}, allowing for significant speedups when ( number of core states ) x ( number of conduction bands ) < ( number of processors )
    • Added n_prinicipal_min, n_principal_max to set the minimum and maximum principal quantum numbers to include in scattering rate calculations involving core initial states.
    • Changed valence -> free parallelization scheme to {i, k}.
    • Changed absorption parallelization scheme to {i, k}
    • Added option to skip saving transition form factors in absorption calculation.
    • Reworked dielectric calculation, added more namelist input options.
    • Changed dielectric parallelization scheme from {i, i'} -> {i, k}, allowing for significant speedups when ( number of valence bands ) x ( number of conduction bands ) < ( number of processors )
    • Added dielectric process which computes just the dielectric.
    • Utilizing type structures to make code more modular/reusable. Each type has at least its own load, save, and print procedure.
      • PW_dataset - handles the plane wave Bloch wave function coefficients
      • dm_model - dark matter model parameters
      • material - collection of target material parameters
      • expt - experimental parameters
      • core_electron - Core electron configuration and STO wave function coefficients for the core electrons
      • and more!
    • Output scattering rate units are now cm^(-2), for easier conversion to cross section constraints.
    • Output absorption rates are total rates given use specified experimental masses and exposures.
    • EXCEED-DM version is now written to output for easy comparison with previous and future versions.
    • Added k_ subgroup to Bloch coefficient data structure for easy access to the wave function coefficients at a given i, k.
      • Updated example files in examples/dft accordingly.
    • Updated core electron configuration file specification
      • Updated example files in examples/(Si, Ge)/core accordingly.
      • Updated utilities/create_elec_config.ipynb.
    • Added example input files for Germanium.
    • Improved/standardized output printing with an info_messages module.
    • Improved comments inside the code, many variable explanations have associated LaTeX'ed equations which can be read by viewing the documentation in a browser.
    • Major update to documentation. Check it out here folder.
    • All modules, procedures, and types have some documentation.
    • Specific documentation pages for all input and output files.
    • /examples/Si/dft/Si_2x2x2_AE_spin.hdf5 are now realistic spin dependent wave function coefficients for Si (just spin-independent ones doubled.).
    • Added preliminary logo, docs/media/exdm-prelim-logo.png.

v0.2.4

  • Reworked the implementation of the vector/pseudoscalar DM absorption calculation.
  • Fixed bug in absorption rate calculation introduced in v0.2.3 which overwrote the main processors transition form factors when computing the self energies.

v0.2.3

  • Calculation of the dielectric for targets with spin-dependent wave functions is now supported.
  • Added routine to time the dielectric calculation.
  • Added spin_degen which accounts for the spin degeneracy factor of the initial (valence) states.
  • Added scalar_LO absorption calculation mode to just compute the leading order contribution for scalar DM absorption.
  • Generalized vector/pseudoscalar DM absorption calculation for anisotropic targets.
    • Self-energy, Pi_vi_vj, is now computed along with the other self energies.
  • Implemented routine to rigorously find maximum magnitude of q for which an FFT will give consistent results across meshes. See find_q_max_FFT routine in FFT_util.
  • Updated install.md

v0.2.2

  • Improvements to absorption module.
    • Initial support for spin-dependent wave functions.
      • More general transition form factors can be computed.
    • Parallelization of velocity integral
  • LAPACK and BLAS are now required. Additional installation instructions have been added to the ubuntu-gnu build.
  • q_s_FFT has been removed as a numerics input option. This is now computed directly with LAPACK routines.

v0.2.1

  • Added a dielectric module which computes the dielectric in the scattering kinematic regime.
    • To screen the rate with a numeric dielectric model, set screen_type = numeric.
    • If load_dielectric_from_file = .FALSE., the dielectric will be computed from scratch. Otherwise the screening factor will come from the dielectric matrix in the dielectric_filename file.
    • Variables relevant for the calculation are loaded through the dielectric namelist. Check out examples in the examples/ folder.
    • Note: the dielectric here is only used to screen. Currently only v -> c transitions are included in the loop. For now, spin-indpendent wave functions only.
  • Added ubuntu-gnu-debug build mode to debug with.

v0.2.0

  • Calculation of dark matter (scalar, pseudoscalar (axion-like), vector) absorption on electrons!
    • Compute the absorption rate, self-energies, and generalized crystal form factors in the absorption limit.
    • Setting process = 'absorption' in the input file switches the calculation to absorption mode. Check out examples in the examples/ folder.
    • Set a variety of electron lifetime/width parameters, width = min( a + b omega , width_max )
    • See https://arxiv.org/abs/2106.12586 for details of the formulation.
    • (Currently only spin-independent, valence -> conduction transitions are supported.)

v0.1.3

  • Partial implementation of general transition form factor in module transition_form_factor.
    • Compute scattering rates for interactions that depend on electron spin with spin-dependent electronic wave functions.
    • Support for valence -> conduction transitions.
    • Note : default is to compute spin-independent scattering rates
  • Added simple testing routines which will help make sure new additions do not break old functionality.
    • Run test with FoBiS.py rule -ex tests.
    • To add new tests just add files to the list inside tests/run_tests.sh
  • Updated documentation

v0.1.2

  • Added functionality for spin dependent (two component) wave functions!
    • Automatically detect whether input DFT data has spin dependence.
    • Valence -> conduction transition rates can be computed with spin dependent wave functions.
    • Added example input file for Si which has spin dependent wave functions.

v0.1.1

  • When timer = .TRUE. a smaller version of the program will be run before the main program, and an estimate of the run time of the full program will be printed.
  • Improved output printing.
  • Updated Si example input files to use file paths relative to the main folder.

v0.1.0

  • Initial beta release.