SFINAE now works for non-const vector without explicit specialization.
Merged read/write basic functor relying heavily on Hana. Tried constexpr if.
Some amendments, partly in try of C++14
Everything migrated, i/o doesn’t work yet.
Atomic can be implemented with constexpr if
This didn’t solve the problem, either.
handleDelimiter implemented
Progress in reading.
Input now works, everything goes with strings instead of char*
Queue is more natural for autostop than list.
Updated logo
Almost there – remaining problem: cpypyqed linked against boost_python instead of boost_python3
Merge branch 'tupleMigration' into Development
Fixes a small bug in mode::AveragedMonitorCutoff
Introspection (is_container) now based on Hana
Update to some of the scripts to make them compile again
hana::tuple output now works
little nicety
serialize_tuple.h inserted into ContainerIO.h
Improved tuple initialization
Updated a few more lingering copyright notices.
String trimming facility with primitive signature to implement liberality with whitespaces in delimiter input.
Necessary small changes in infrastructure for QbitModeIntelligentEvolution
Merge branch 'python3migration' into Development – only partly successful feature branch
std::tuple everywhere now
Updated copyright notices to 2020.
Bug with formdouble::zeroAdditional
templatized evolution::Pars
2017 => 2020 in copyright statement
Overall precision observed in the commented-out parts of output as well.
The comment character issue is solved, but we would still need the option for formatting the output into CSV
Better integrated the QbitPhaseNoise bundle into the Qbit bundle
Better integrated the QbitPhaseNoise bundle into the Qbit bundle
Better integrated the QbitPhaseNoise bundle into the Qbit bundle
Better integrated the QbitPhaseNoise bundle into the Qbit bundle
Better integrated the QbitPhaseNoise bundle into the Qbit bundle
This could be done in the python wrapper of evolve only (so as not to make the main part of the framework depend on Boost.Python) by letting the trajectory write on a stringstream and using numpy::loadtxt to load from the string into an ndarray.
This could be done in the python wrapper of evolved only (so as not to make the main part of the framework depend on Boost.Python) by letting the trajectory write on a stringstream and using numpy::loadtxt to load from the string into an ndarray.
This could be done only in the python wrapper of evolved only (so as not to make the main part of the framework depend on Boost.Python) by letting the trajectory write on a stringstream and using numpy::loadtxt to load from the string into an ndarray.
Trajectory output to be returned as boost::python::numpy::ndarray
Little cosmetics to StateVector.addTo (including perhaps the elimination of a bug?)
Trajectory-bundle rationalized with base-class chaining
trajectory::Ensemble & quantumtrajetory::EnsembleMCWF rewritten
Deprecated master_fast
Rationalized some naming in trajectory::Ensemble
Eliminate a few errors with gcc 5.x and clang 4.x
Added a move constructor to StateVector and DensityOperator
Specialization of blitzplusplus::unaryView and binaryView for RANK=1 & 2, respectively
Corrected error in HomodynedMode
alternative L*rho*Ldagger calculation
Achieved on commit #6c1188
Implementation for finite-temperature mode => speedup by factor of 1.9
Got away with the strategy-madness in multilevel::Liouvillean in favor of the base-class chaining idiom.
Got rid of obsolate file-type specification for emacs
Migrated most usecases
And, finally, the implementation for MultiLevel.
Streamlined i/o of tuples and tiny vectors.
Reverted previous patch and eliminated the set flag + exception in Trajectory.h
Implementation for BinarySystem
Tried to patch error. Some exceptions still thrown during check_full.
Generalized NX_CoupledModes to cover also XX and XX in RWA
Introduced a generic type-wrapper for the MultiLevel parameter storage.
Small in the wake of CommentingStream
Merge branch 'JrhoJdagger' into Development
Merge branch 'CommentingStream' into Development
First implementation for Mode => speedup by factor of 5/3
Reintroduced abs-value accidentally removed in 10e6c3b4
Basic infrastructure for SuperoperatorCalculation ready.
Error spotted by new version of clang (integer not implicitly convertible to bool @ compile time?)
And finally the implementation for composites
Fundamental infrastructure
Some template typedefs
Almost there. Remaining problem: MCWF_Trajectory may experience dpLimit overshoot at the very onset, which it may want to log. At this point the logstream is yet unset, so this results in an Exception.
Homodyned – error in implementation
Migrate to python3
Migrate to python3
Eliminated exception specifier (deprecated in c++11)
Almost there – remaining problem: cpypyqed linked against boost_python instead of boost_python3
This didn’t solve the problem, either.
Eliminated superfluous virtual specifications from final overrides.
Eliminating a few warnings and an error with newer compilers
Adaptive low-rank approximation and denoised Monte Carlo approach for high-dimensional Lindblad equations
It looks like Eigen is also developing a multi-array template that they call tensor: http://eigen.tuxfamily.org/dox-devel/unsupported/groupCXX11TensorModule.html Looks promising…
Replace blitz::Array with boost::multi_array or, MAYBE/RATHER eigen::tensor
2016 => 2017 in copyright info
Merge branch 'master' into Development
Updated testsuite and some more modules in the wake of 8bd835f8
LossySpinWithout => LossySpinSch
Removed unused variables & parameter
Factored out logging pertaining to the coherent step from MCWF_Trajectory to Evolved/Adaptive,
Patching a compilation error with g++ 6.3.1
alternative L*rho*Ldagger calculation
Possible clash between argparse and our ParameterTable
Updated year in copyright info
Merge branch 'master' into Development
Updated year in copyright info
Merge branch 'master' into Development
Added a list of publications using/citing us to...
load_jumps function for tools.io
Disable writing of compressed state files.