UKRMol+: UKRMol-out

Outer region programs for the re-engineered UK computational implementation of the R-matrix method for the treatment of electron and positron scattering from molecules (BTO/GTO continuum). Also calculates photoionization cross sections.

Docker container with UKRmol+ 3.3.0 (GBTOlib 4.0.0. + UKRmol-in/out 3.3.) is available from Docker HUB: 

• For instructions on how to use follow the documentation on Docker HUB
• Obtain the Docker container using: `docker pull zdenekmasin/ukrmol_plus:3.3.0`

New features of this version

• Effective core potentials implemented

• One-photon ionization delays

• Generalize multidip_rabitt_delays to higher orders

• Calculation of aligned distributions

• Support for calculation of Aligned DCS and Molecular Frame DCS

• Calculate complex beta parameters for RABITT

• Allow unequal XUV and IR polarizations in RABITT

• 2-photon beta parameters for arbitrary polarized RABBITT

• Calculate also emission-integrated molecular-frame RABITT

• Per-partial-wave asymptotic corrections to RABITT delays

• Allow passing input file to MPI_RSOLVE on command line

• Consider complete bound state wave-function in MULTIDIP

• MULTIDIP: user-input of lab-frame polarisation and saving of 1-photon dipoles in RSOLVE format

• Accelerate R-matrix computation on GPU

Bug fixes

• Do not inflate minimal lm_rg in RMT_INTERFACE

• Extend width of energy index to 6 characters

• Fix BLAS integer kind in multidip_io::load_mapped_matrix

• Avoid truncating high-dimensional integrals in MULTIDIP

• Skip COMPAK in MPI_RSOLVE if no inner region dipoles available

• Correct overall sign of multi-photon cross section

• Support fixed property formats in COMPAK (and PHOTODELAYS)

• Correctly type variable KNOTS in RSOLVE

• Fixed a missing variable initialization

• Fix NaN replacement at first energy in DIPELM

• Fix indexing in a_re_sp_harm resulting in wrong sign of some spherical harmonics in DIPELM and MULTIDIP

• Fix invalid memory accesses in RESON

Optimizations of speed and memory use

• Improve MPI_RSOLVE memory scaling

• Performance and other improvements to MULTIDIP

• Use real arithmetic when calculating T-matrix in MULTIDIP

• Use closed-form expression for ion-ion integral in MULTIDIP

Other improvements

COMPILATION/INSTALLATION/DOCUMENTATION

• Automatically assume ScaLAPACK also with -qmkl-ilp64=cluster

• Use OpenMP flags when linking multidip_rabitt_delays

• Restore GCC 8.x compatibility

• Simplify build & install path handling in CMake scripts

• Do not require C++ in CMake scripts

• Lower required version of CMake from 3.14 to 3.5

• Avoid CMake 3.10 deprecation warning

• Rename libutil to libouterutil to avoid clash with libutil from GLIBC

• Minor changes to Doxygen documentation

• Include "python_read.html" in DIPELM doc

TEST SUITE

• add PCCHF tests for electron scattering

• Update test suite after switching off nuclear properties in photoionization

• Update of output files for positron tests

• Copy photoionization test outputs to "outputs" subfolder

MULTIDIP

• Implement long-range amplitude correction to cc delay

• Get also kinetic energy from Ak file when reading it

• Read raw dipoles filenames from files, not from command line

• Optionally cache outer integrals 

• Use raw MPI in place of Fortran coarrays

• Improve integral precomputing

• Abort on invalid charge

• Clarify how the initial state is chosen

• Remove unused variables

• Use rasym = 100 as default

• Cope with NaNs in  smoothing

• Change default for WITH_MMAP to ON 

• Make ion_ion_analytic=.true. the default 

• Calculate also emission-integrated molecular-frame RABITT

OTHER

• Require equal CALCAK and CALCDIP(3) when both present in RSOLVE/MPI_RSOLVE

• Inclusion of a cleaned-up COULCC in SIEGERT ???

• Add atomic hydrogen (RABBITT) continuum-continuum delay to PHOTODELAYS

• Assume Hartree units on input to PHOTODELAYS

• Increase number of states which can be used in SWINTERF and TMATRX

• EIGENP IWRITE and LUPHSO consolidation

• Remove some debug printouts from BOUND

• Avoid hardcoding units 5 and 402 in IXSECS

This version uses GBTOlib 4.0.0.