ChangeLog

2.2.0 relative to 2.0.6
=======================
- Parallelization using MPI. The diagonalization can now be spread
  over several calculation nodes. In combination with the SMP
  parallelization in LAPACK routines, even higher performance can now
  be achieved.
- Portability improvements (for example, the code now compiles under 
  Mac OS X without any tweaking).
- Calculation of the temperature-dependent conductance G(T), and 
  of the temperature-dependent moments (cf. specgt, speci1t, speci2t).
- Optimized routines for spectral function calculation, contributed
  by Markus Greger.
- Enhanced support for "global" operators. A number of such operators
  are defined for a number of different symmetry types. These can
  be used to calculate the temperature-dependent susceptibility
  for the general case (specchit).
- New symmetry types: SU(2)_iso x SU(2)_iso x U(1)_spin,
  SU(2)_iso x SU(2)_iso, U(1)_ch x U(1)_ch x U(1)_ch (for three-channel
  problems), U(1)_spin x Z_2
- Improvements in output from the code, diagnostics, etc.
- Calculation of an estimate of the Kondo temperature for the Kondo
  and Anderson models (TKW output line).
- Output of real and imaginary parts of spectral weights (reim).
- Extrapolation of thermodynamic calculations to high temperatures
  (tdht).
- Tools for smoothing of spectral functions (intavg) and for smoothing
  the thermodynamical quantities (tdavg).
- Adaptable broadening tool (bw), implementing the method introduced
  by A. Freyn and S. Florens.
- Old docs, tests and examples were removed. The documentation will
  be distributed separately.
- Some defaults for parameters have been changed, taking into account
  recent algorithmic developments and the accumulated experience with
  running the NRG calculations.
- Some bugs fixed (for instance, spectral function calculation
  for symtype=NONE).
- Huge library of automatic regression tests (distributed separately,
  big file!).
- "Decoupling tools" (matrix, diag, unitary in tools/dec). 
  It is now possible to run NRG Ljubljana without Mathematica by using 
  perl scripts to generate the input 'data' file from a template file 
  (generated using option=TEMPLATE).

2.2.1 relative to 2.2.0
=======================
- Output of complex expectation values (for non-Hermitian operators)

2.2.2 relative to 2.2.1
=======================
- Bug fix in CFS code
- smooth=newsc broadening scheme for problems with a superconducting gap
- Code cleanup, some fixes in the README file
- Schur decomposition is now the default way of diagonalizing the
  Hamiltonians. The old behavior can be restored using the NOSCHUR option.

2.2.3 relative to 2.2.2
=======================
- Bug fix in CFS code for complex problems

2.3
===
- Full density matrix code (for spectral functions and expectation values)

2.3.9
=====
- 2x2 spin structure for symtype=U1

2.3.15
======
- store=one vs. store=split

2.3.16
======
- "general approach" for computing the linear conductance at finite
temperatures
- binary data now contain complex values if P::reim=true

2.3.20
======
- Compilation and installation of tools is now optional. Use
"./configure --with-tools" to enable.
- orthogonality check for checkdiag=true

2.4.0
=====
- Backwards compatibility breaking; remove cruft from the past.

2.4.0.1
=======
- Drop F77 requirements. No BLAS/LAPACK autodetection.

2.4.0.3
===========================
- CFSold, CFSgt, CFSls

2.4.0.4
===========================
- different handling of U, Gamma, delta, t parameters

2.4.1.0 relative to 2.4.0.4
=========================
- Interleaved Wilson chains (substeps). Presently implemented for QS and SPSU2 symmetry types.

2.4.1.1 
=======
- fixed gather* script, added some "use strict" statements

2.4.1.2
=======
- Substeps for SPU1

2.4.1.3
=======
- Gaussian broadening in broaden tool

2.4.1.4
=======
- U_d, D_d, CU_d, CD_d, AU_d, AD_d operators

2.4.1.6
=======
- Improved compatibility with recent C++ standards [unordered_map,
headers]

2.4.1.7
=======
- sign: int -> double in spec_*.cc

2.4.1.8
=======
- my_warning("tr1=") bug fix

2.4.1.9
=======
- offset by 1 error in nrgchain2x2 fixed

2.4.1.10, 12.1.2015
========
- HAVE_BOOST_MPI -> NRG_MPI

2.4.1.11, 31.1.2015
========
- bigfix for ireducsigma[] - spherical tensor operator

2.4.1.12, 6.3.2015
========
- broaden - support for accumulation point away from omega=0
- bins.h - support for accumulation point away from omega=0

2.4.2, 11.8.2015
=====
- QST basis for 3-channel problems

2.4.2.1, 25.8.2015
=======
- specmoments tool

2.4.2.2, 24.9.2015
=======
- SPSU2LR symmetry type

2.4.2.3, 25.9.2015
=======
- versioned binary number, to avoid NFS issues

2.4.2.4, 13.10.2015
=======
- 3-channel SPSU2 code
- removed doublet-self code

2.4.2.5, 22.10.2015
=======
- SPSU2T code

2.4.2.6, 30.10.2015
=======
- QSC3, SPSU2C3 codes
- complex correctness in initial.m

2.4.3.0, 3.11.2015
=======
- refactorization of symmetry codes
- bug fix for line by line sending of matrices over MPI

2.4.3.1, 6.11.2015
=======
- bug fix in calc_abs_energies (RMAX -> NRSTATES)

2.4.3.3
=======
- bug fix for "td" in ISOLR

2.4.3.4
=======
- orbital-triplet creation op for QST
- d[] -> d in operators.m, _Function -> _ in initial.m

2.4.3.5
=======
- QST troubleshooting

2.4.3.6, 25.2.2016
=======
- QSTZ symmetry type

2.4.3.7, 2.3.2016
=======
- QSZTZ symmetry type

2.4.3.8, 7.3.2016
=======
- QJ symmetry type
- mtDoubletOp[]
- some reshuffling in operators.m, lops now sorted alphabetically

2.4.3.9, 17.3.2016
=======
- 4ch QS code

2.4.3.10, 24.4.2016
========
- support for numerical coefficient tables (see QST)

2.4.3.11, 14.6.2016
========
- global magnetic field

2.4.3.12, 24.6.2016
========
- POLARIZED for symtype=NONE

2.4.3.13, 21.9.2016
========
- reorganized some code in adapt
- removed dmft/ since it is obsolete
- removed rescaling of xi() in data file

2.4.3.14
========
- support for automatic rescaling (bandrescale parameter)

2.4.3.16
========
- doublet spectra for QJ

2.4.3.17
========
- broaden tool can do a final Gaussian broadening pass

2.4.3.18, 13.10.2016
========
- fixes in spectrum calc for QJ and QST

2.4.3.19, 19.10.2016
========
- fixes in quadrupole operator handling

2.4.3.20, 21.11.2016
========
- spin and orbital triplet operators for QST

2.4.3.21, 1.12.2016
========
- bugfix in offdiag_build(): factor -> factor_rescale

2.4.3.22, 19.1.2017
========
- substeps for QSZ

2.4.3.23, 19.1.2017
========
- LRSPIN option

2.4.3.24, 25.7.2017
========
- SYMTYPE=P (total fermion number parity conservation)

2.4.3.25, 30.8.2017
========
- SYMTYPE=PP (separate fermion number parity conservation in two separate channels)

2.4.3.26, 13.11.2017
========
- support for channel-mixing Wilson chains (rungs=true) for QS and QSZ

2.4.3.27
========
- hilb tool for converting im/resigma -> im/reaw

2.4.3.28
========
- support for complex Wilson chain coefficients in initial.m (manual
and manual_V), function ImportTable[]

2.4.3.29
========
- improvements in hilb tool (imsigma -> imaw, Steffen interpolation)

2.4.3.30
========
- BAND=manual_nambu

2.4.3.31, 14.8.2018
========
- bug fix in the coefficient table for symtype=P for doublet operators

2.4.3.32, 7.1.2019
========
- cosmetics in initial.m (basops etc.)

2.4.3.33, 29.1.2019
========
- portability fix in calculate_Z, removed boost::lambda::bind dependence

2.4.3.34, 28.3.2019
========
- VDIM

2.4.3.35, 27.6.2019
========
- P::ZBW, fixes in calc_generic_DMNRG(), mergeNN2half(), LAST_ITERATION(),
window_function()
- new variable Nlen, changes in determine_Nmax()
- new function do_no_recalc(). Fixes in do_recalc_kept(),
do_recalc_all()
- new function doZBW(), changes in nrg_after_diag()
- changes in start_calculation(). ?? fdm density matrix handling
- additional nrglog lines for log=@
- C++14 constructs, for loops.
- time "spec" as a whole, not separately for different spectra types
- removed all broadening schemes except "new". The idea is that
nowadays we always do broadening using an external tool.
- FDMold, CFSold removed
- get_loop_dim() moved to spec.cc

2.4.3.36, 12.7.2019
========
- Bins::setlimits, P::emin, P::emax, code cleanup in bins.h
- sign fix for CFS calculations for bosonic operators to bring it in
line with the standard convention (and with the FT, FDM and DMNRG
implementations, which already had correct sign for bosonic operators)

2.4.3.37, 19.7.2019
========
- more for() loop modernisation
- cleanup, remove functions that never got called
- removed value_raw(), replaced by shift. Cleaned up code handling absenergy.
- getNz script
- style fixes, removed empty lines in nrg.cc (TO DO: other files)
- modernized regression testing infrastructure
- all output to cout, nothing goes to cerr (this makes problems easier to
spot in the logs)
- removed some boost libraries, switched to C++11 solutions
(unordered_map, lambdas)

2.4.4.1, 22.7.2019
=======
- P::finitemats, P::fdmmats, etc., P::mats
- removed NRDOTS from data file, removed P::dots
- removed P::T0opt
- matsubara.h
- BaseSpectrum cleanup, now contains all relevant data field, fullname()
- SpectrumMatsubara, ChainSpectrumMatsubara
- cleanup of open_files_spec, new open_files()
- enum class matstype (fermionic, bosonic)
- enum class axis (RealFreq, Temp, Matsubara)
- matstypestring, axistypestring
- changes in spec.cc: creation of ChainSpectrum in a switch statement, 
relevant parameters now read from BaseSpectrum within calc_generic()
- OPENSPEC is now a function. 

2.4.4.2, 23.7.2019
=======
- cleanup of calc_generic(), BaseSpectrum passed as parameter in place of sign

2.4.4.3, 23.7.2019
=======
- DMNRGmats, tested & works!
- *SpectrumMatsubara2

2.4.4.4, 24.7.2019
=======
- removed get_loop_dim(), dim* parameter to calc_spec functions
- x.begin() -> begin(x)
- functors replaced by lambdas
- removed recalculation wrappers
- myrank defined for non-MPI case
- new safeopen()
- new output_val()
- calc_generic3()
- major cleanup of diag.h, removed unnecessary tests, removed macros

2.4.4.5, 25.7.2019
=======
- removed masquerading in Spikes
- fixed initializations in classes
- int,unsigned int,etc. -> size_t where used as index in for loops, or
as size or counter variable. Where appropriate, uninitialized value
changed from -1 to 0.
- renamed nrg-invar-ANY.cc to invar.cc
- NRG_WORKDIR from environment, creates a uniquely named work subdirectory
- reports unused parameters, parameter list is sorted, parameters that were
modified from the default values are indicated by an asterisk

2.4.4.6, 25.7.2019
=======
- calc_generic3_FT_v3mm() implemented

2.4.4.7, 26.7.2019
=======
- SPECTYPE converted to a pointer to object (class SPEC), use
polymorphisem to create appropriate ChainSpectrum and to call
appropriate calc() member

2.4.4.9, 1.8.2019
=======
- source reorganization for DM calculations
- alternative calculation for FDM on Matsubara axis (omega=0 for
bosonic case not handled + it is slower, thus commented out,
but in principle it works)

2.4.4.10
?

2.4.4.11, 4.9.2019
========
- new implementation of FDM Matsubara based on automated compter-algebra
derivation; it also is faster since it avoids redundant calculations.
- dropped version number (2_4) from filename
- change in default path for modules in nrginit
- some obsolete scripts removed from the distribution

2.4.4.12, 5.9.2019
========
- v3mm implemented for FDM
- total weight reported during Matsubara ChainSpectrum merging
- some const correctness fixes
- fixed Nlen handling for ZBW case, as well as setting N to Ninit (and
scale according to Ninit-1)

2.4.4.13, 12.9.2019
========
- optimizations in FDM_v3mm
- clean up warnings from clang
- comment out assertions in Matsubara2::add
- size_t correctness

2.4.4.14, 13.9.2019
========
- N is now size_t

2.4.4.15, 14.9.2019
========
- nrg_diagonalisations_serial() removed
- OpenMP parallelization for Matsubara calculations (FT and FDM)

2.4.4.16, 17.9.2019
========
- optimizations for FDM calculations of 3-pt vertex functions

2.4.4.17, 18.9.2019
========
- fixed a bug introduced in nrg.cc (call to shift_abs_energies) in
2.4.4.16, reverting to previous version

2.4.4.18, 19.9.2019
========
- bug fixing in 2-pt GF calculation with FDM algoritm (new implementation)

2.4.4.19, 20.9.2019
========
- increased the default precision of output by 4 decimals (6 -> 10)
- LOOP_D and LOOP_K macros, ENG(x,matstype) in spec_FDM.cc
- code cleanup in spec_FDM.cc

2.4.4.20, 20.9.2019
========
- FDMmats optimization

2.4.4.21, 24.9.2019
========
- spec_FDM.cc cleanup, ENG() removed, cleared up the mess with
transpositions

2.4.4.22, 24.9.2019
========
- new implementation of v3mm FDM
- v3mm FT removed (last tested working version = 2.4.4.20)

2.4.4.23, 25.9.2019
========
- P::diagth
- v3mm FDM finished 

2.4.4.24, 18.10.2019
========
- more detailed logging in matrix.cc
- bug fix in sym-U1.cc (incorrect handling of zeta(3,4) coefficients
  in the case of complex coefficients)
- reverted: sym-U1.cc was fine, initial.m had to be fixed instead

2.4.4.25, 22.10.2019
========

TO DO:
- TD-NRG
- parity projected calculations