For more detailed information about the changes see the history of the
`repository <https://github.com/votca/xtp/commits/master>`__.
Version 2021.1 (released 18.07.21)
==================================
Version 2021 (released 13.03.21)
================================
- fix CI on rawhide (#663)
- add xtp_qmmm2qm to cmake install (#665)
- add auto-cancel workflow to GitHub Actions (#674)
Version 2021-rc.1 (released 15.01.21)
=====================================
- G0W0 rpa energies are fixed in BSE (#383)
- Hqp part of BSE can be separate from QP range (#388)
- Fixed Cuda implementation (#391)
- Plain, linear, and Anderson mixing in evGW
- Use a general syntax to represent all QMPackages input (#318)
- Remove support for both Gaussian and NWChem (#318)
- Fixed executable path check (#400)
- Usage of offdiagonal elements of Hqp in BSE optional, default: with
offdiagonals (#402)
- refactored MO reordering of external QMPackages (#406)
- Add defaults to XML (#319, #413)
- fix build with gcc-10 (#422, #424)
- Add name command line argument (#426)
- fix usage of PROJECT\_SOURCE\_DIR in cmake (#431)
- check that temperature is positive for KMC (#432)
- document cmake options and packages (#434)
- remove long gone MKL option for gitlab-ci (#435)
- add documentation using Sphinx (#433, #615)
- add kokkos-based Ewald solver (#444, #446, #485)
- allow multiple choices for the calculator options (#442, #445)
- move bsecoupling defaults to xml (#451)
- address missing includes (#452)
- fix readjobfiles path (#453)
- standardising include order and style in header files (#448)
- changed to OpenMP reductions instead of hand crafted solutions (#466,
#471, #493)
- switch to GitHub Actions as CI (#467, #474, #477, #482, #500, #502,
#518, #549, #637)
- move tutorial to its own repo (#450)
- removed share/packages folder (#475)
- copy orca \*.gbw files in eqm (#472)
- improve mkl support (#479)
- add def2-tzvp, def2-tzvpp, def2-qzvp, def2-qzvpp for Rb,I,Ag,Xe
(#480)
- create a map between orca and libxc functional names (#461)
- fix path to share/data (#486, #487)
- remove special XML symbols (#488)
- standardize includes in src files (#462)
- add cmake exported target (#489)
- update codacy badge (#490)
- add mkl builds to CI (#496)
- convert markdown to rst (#497)
- enable CXX only in CMake (#499)
- implement RPA correlation energy (#498)
- added verbose option for rootfinder (#503)
- fix orca pointcharges (#504)
- remove constant cutoff from neighborlist (#508, #509)
- fix default and flags properties issue (#511)
- add more checks to the dft_parse_part (#510)
- add NOSCF option for iqm in internal and external DFT (#483)
- added versions to output files (#523)
- write an empty state file if it does not exist yet (#526)
- fix double unit conversion (#531)
- Unittests can read input and reference values from files (#529)
- fix double SCF in ORCA keyword line (#533)
- xtp internally uses only single shells. (#539)
- Refactor orbital reordering (#537)
- fix build with clang-11 (#546)
- Add a molden parser to xtp (#547)
- Correct self-energy frequency derivative (#550)
- fix multiline orca keyword (#552)
- Contour Deformation Approach (CDA) for self-energy implemented (#4, #548)
- BSE with perturbative dynamical screening (#3, #560)
- Correction of full BSE energy contribution in analysis output (#560)
- Bump required CMake version to 3.12 (#568)
- Refactored calculator interface (#570)
- Fixed boost test deprecation warnings (#569)
- Refactored CUDA-OpenMP integration (#571)
- replaced own integrals with the libint library (#559)
- improved kmc output (#574)
- fix build on Ubuntu 16.04 (#577)
- use unique_ptr in factory (#581)
- fix QMMM with QP (#590, #591)
- fix bug in setCenter of AODipole (#593)
- extra check in orca unit test (#594)
- fixed atomid numbering while adding containers (#599)
- added tool for automatic mapping file generation (#595)
- Added GPU support for RPA (#601)
- fixed int64 interface to RDKIT (#602)
- remove kokkos dependent ewald (#612)
- fixed Sphinx warnings (#609)
- better error for exceeded angular momentum in libint (#613)
- fixed wrong QP-QMMM energy with non-zero QPmin (#620)
- fixed density analysis options (#627)
- Use libint transformation matrices (#630)
- Added script to extract orbitals from qmmm statefile (#629)
- clean up XTP app message (#631)
- fixed EXX 3c mode (#640)
Version 1.6.4 (released 12.01.21)
=================================
- fix build on openSUSE (#622)
- refactored logger (#625)
Version 1.6.3 (released 09.12.20)
=================================
- switch to ghcr.io for CI (#555)
- fixing Gaussian guess read keyword (#562, #563)
- improved kmc output (#575)
- use master .clang-format in format action (#582, #584)
- strip windows line endings in readers (#596)
Version 1.6.2 *SuperGitta* (released 22.08.20)
=================================
- move CI to GitHub Actions (#512, #514, #516,
#519)
- add more checks to the dft_parse_part (#521)
- make unit tests work in parallel (#530)
Version 1.6.1 (released 21.06.20)
=================================
- fix warnings on Ubuntu 20.04 (#438, #460)
- remove obsolete boost define (#440)
Version 1.6 *SuperPelagia* (released 17.04.20)
==============================================
- fix 32-bit build (#381, #380)
- remove duplicated basissets (#384, #386, #387)
- fix clang-10 warnings (#394)
- fix unit\_test\_eeinteractor on OpenSUSE (#341, #428)
Version 1.6\_rc2 (released 10.02.20)
====================================
- fix remove giant logo from tarball (#337)
- fix assertions related to GLIBCXX\_ASSERTIONS (#345)
- remove unused boost serialisation (#346)
- fix build on 32-bit archs (#347)
- add ENABLE\_HIGH\_MEMORY\_TESTS cmake option (#356)
- fix copyright (#363)
- remove old doxygen target (#365)
- fix some gcc10 warnings (#376)
- Add external fields to dft calculations (#351, #353)
- added def2 and cc basis sets (#355)
- added apdft (#350)
- added test to cubefile reader (#344)
- fix state tracker (#333)
- grid class refator (#335)
- changed ppm screening (#371)
Version 1.6\_rc1 (released 04.12.19)
====================================
- completely new statefile in hdf5 format for larger systems
- new electrostatics with PCG solver
- new QM/MM engine with freely configurable regions
- exact GW-BSE for small systems
- new iterative matrix solvers for large systems
- CUDA support for parts of GW-BSE
- full LAMMPS support
- improved testing
- new tutorial with more functionality
- deleted netbeans support
- removed old manual
- Major CMake refactor
Version 1.5.1 (released 20.11.19)
=================================
- remove exit() calls in the library
- fix build on CentOs7
Version 1.5 *SuperVictor* (released 31.01.19)
=============================================
- enable gitlab CI
Version 1.5\_rc3 (released 19.01.19)
====================================
- travis: fixed bug in building tags
Version 1.5\_rc2 (released 16.01.19)
====================================
- fix parallel build of manual
- fix usage on inkscape on arm arch
- clean up namespace in header
Version 1.5\_rc1 (released 28.12.18)
====================================
- optimized GW-BSE code and integral engine
- added closed shell DFT code which supports hybrid functionals
- removed ctp dependency
- atm no support for site energy calculation
- CHELPG fit for ground and excited states
- merged igwbse and idft into one calculator
- added unit and integration test
- improved geometry optimiser
- replaced ublas with Eigen3
- replaced boost serialisation with hdf5 files
Version 1.4.1 (released 02.09.17)
=================================
- fix pkg-config file
Version 1.4 (released 29.10.16)
===============================
- fixed a bug in gwbse
- added missing copyright
- cmake: fixed underlinking
Version 1.4\_rc1 (released 26.09.16)
====================================
- include manual
- an extension of the whole workflow from: electrons and holes, to
singlet and triplet excitons
- a fully functional GW-BSE code optimized for: molecular properties,
including excited state geometry optimizsation
- Inclusion of LIBXC to calculate Exchange correlation matrices
- allowing interfacing GW-BSE with many quantum mechanical packages
- support for ORCA DFT package
- framework to use stochastic models to generate larger system
- better performance of larger systems
- new calculators: egwbse,igwbse,ewald,.....
- support for intel mkl library and compilers for better performance
- A periodic polarisation embedding: to calculate classical
configuration energies without cutoffs
- xtp\_update\_exciton to update state file to newest format
- integration of moo and kmc into xtp for easier installation
- kmc\_lifetime calculator to simulate exciton movement with lifetimes
- partialcharges to extract atomic charges from qm calculation
- renaming from ctp to xtp
- many bugfixes
Version 1.3 (released XX.09.15)
===============================
- new executables: ctp\_tools, ctp\_dump, ctp\_parallel,
xtp\_testsuite, xtp\_update
- ctp\_tools wraps light-weight tools that assist e.g. in generating
the system mapping file
- ctp\_dump extracts information from the state file to human-readable
format
- ctp\_parallel wraps heavy-duty job-based calculators: allows
synchronization across processes
- ctp\_testsuite provides an easy-to-use environment to run: selected
tests, individual calculators
- ctp\_update updates an existent state file to the current version
- new calculators: edft, idft, pdb2map, xqmultipole, ...
- edft / idft: provide interfaces to the GAUSSIAN, TURBOMOLE & NWCHEM
package, using packages computes: couplings, internal energies,
partial charges
- pdb2map (generates a system mapping file from an input coordinate
file)
- xqmultipole computes classical configuration energies of: charged
clusters embedded in a molecular environment
- enhanced usability via the command-line help, tutorial & test-suite
- a GUI tutorial assists with the first practical steps in using
VOTCA-CTP
- an extended and homogenized help system provides: short infos on
individual calculator options from the command line
Version 1.0 (released 23.10.11)
===============================
- parallel evaluation of site energies using: Thole model + GDMA -
Tinker no longer required
- much clearer input files (and many more checks for input errors)
- most of calculators are parallel and can be used on a cluster
- bug in zindo/ctp interface fixed
- state file now contains: the atomistic trajectory, rigid fragments,
conjugated segments
- support for several MD frames