910
Comment:
|
← Revision 5 as of 2009-08-19 13:05:45 ⇥
3614
|
Deletions are marked like this. | Additions are marked like this. |
Line 1: | Line 1: |
## page was renamed from RecentChanges4 = Recent Changes to Dynare version 4 = |
= Version 4.0.4 = |
Line 4: | Line 3: |
== Version 4.0.1 == | * Fixed forecasting after stoch_simul * Fixed crash of the smoother when only one exogenous variable * Fixed issue with compilation of MEX files under 64 bits platforms with Matlab R2009a * Fixed name clash when MOD file is called "ramsey.mod" and computes optimal policy * Fixed USE_DLL option at order 2 * Updates to the reference manual in sections on model resolution and estimation |
Line 6: | Line 10: |
Bugs fixed: | = Version 4.0.3 = * Many bugfixes, including: * fixed estimation of constants in linear models * determistic exogenous now work when there is more than one such variable * fixed recursive estimation * fixed OSR * fix to interface to Global Sensity Analysis toolbox * "simul" and "periods" options are now accepted simultaneously by "stoch_simul" * preprocessor no longer crashes when there is an unexpected EOF * operator precedence is now correct in macro-processor * estimated_params_init and estimated_params_bounds now work in conjunction with correlation of endogenous variables * fixed bug in the macro-processor, when the body of a @#for loop is empty * fixed memory management bug in the preprocessor related to the filename being parsed * fixing bug for mode_compute=5,6 * build_matlab.m script works on Matlab for Windows 64 bits * Structures M_ and options_ are saved in the "_results.mat" output file * Structure bayestopt_ is saved in the "_mode.mat" output file * Added option "notmpterms" to preprocessor, for disabling computation of temporary terms * Macro processor reports the location of @#for or @#if statement when not matched by an @#endfor or an @#endif * Many updates to the reference manual for 4.0 (which is not yet finished) * All code source is now under a free license (GNU General Public License 3 for most of them) * All documentation (including user guide) now licensed under the GNU Free Documentation License 1.3 = Version 4.0.2 = * Fixed bugs in Kalman filter and smoother * Fixed bug with "corr" statement of "shocks" block * Added error message if number of equations differs from number of endogenous variables (except when doing optimal policy computation, or standalone BVAR estimation) * Fixed bugs with "dynatype" and "dynasave" commands * Fixed name collisions with Matlab functions "sec2hms" and "bicgstab" * Fixed some licensing issues * Removed obsolete "dr_algo" option * Fixed crashes under Octave due to graphics limitations * Added error message when model local variables (pound sign expressions) are used outside model scope * Included AIM subdirectory in packaging * Estimation: modified computation of log(det(invhess)) to solve cases in large models where det(invhess) is smaller than machine precision = Version 4.0.1 = == Bugs fixed == |
Line 14: | Line 58: |
Improvements: | == Improvements == |
Version 4.0.4
- Fixed forecasting after stoch_simul
- Fixed crash of the smoother when only one exogenous variable
- Fixed issue with compilation of MEX files under 64 bits platforms with Matlab R2009a
- Fixed name clash when MOD file is called "ramsey.mod" and computes optimal policy
- Fixed USE_DLL option at order 2
- Updates to the reference manual in sections on model resolution and estimation
Version 4.0.3
- Many bugfixes, including:
- fixed estimation of constants in linear models
- determistic exogenous now work when there is more than one such variable
- fixed recursive estimation
- fixed OSR
- fix to interface to Global Sensity Analysis toolbox
- "simul" and "periods" options are now accepted simultaneously by "stoch_simul"
- preprocessor no longer crashes when there is an unexpected EOF
- operator precedence is now correct in macro-processor
- estimated_params_init and estimated_params_bounds now work in conjunction with correlation of endogenous variables
- fixed bug in the macro-processor, when the body of a @#for loop is empty
- fixed memory management bug in the preprocessor related to the filename being parsed
- fixing bug for mode_compute=5,6
- build_matlab.m script works on Matlab for Windows 64 bits
- Structures M_ and options_ are saved in the "_results.mat" output file
- Structure bayestopt_ is saved in the "_mode.mat" output file
- Added option "notmpterms" to preprocessor, for disabling computation of temporary terms
- Macro processor reports the location of @#for or @#if statement when not matched by an @#endfor or an @#endif
- Many updates to the reference manual for 4.0 (which is not yet finished)
- All code source is now under a free license (GNU General Public License 3 for most of them)
- All documentation (including user guide) now licensed under the GNU Free Documentation License 1.3
Version 4.0.2
- Fixed bugs in Kalman filter and smoother
- Fixed bug with "corr" statement of "shocks" block
- Added error message if number of equations differs from number of endogenous variables (except when doing optimal policy computation, or standalone BVAR estimation)
- Fixed bugs with "dynatype" and "dynasave" commands
- Fixed name collisions with Matlab functions "sec2hms" and "bicgstab"
- Fixed some licensing issues
- Removed obsolete "dr_algo" option
- Fixed crashes under Octave due to graphics limitations
- Added error message when model local variables (pound sign expressions) are used outside model scope
- Included AIM subdirectory in packaging
- Estimation: modified computation of log(det(invhess)) to solve cases in large models where det(invhess) is smaller than machine precision
Version 4.0.1
Bugs fixed
- some files are missing in Debian and Ubuntu packages (only affects Matlab users who don't have the Statistics Toolbox)
typo at the top of DiffuseKalmanSmootherH1.m (thanks to Tsu-ting Tim Lin for reporting)
- call to cumulative normal distribution is broken (thanks to Karl Walentin for reporting)
- script for building MEX files under Matlab R2008a is broken (thanks to Slim Bridji for reporting)
- homotopy mode 3 is broken
- BVAR à la Sims: Dynare doesn't accept integer values for tau and lambda parameters (thanks to Aqua for reporting)
Improvements
- new solve_algo=4 for computing steady state: similar to solve_algo=2, except that it behaves differently when the Jacobian becomes nearly singular or badly scaled
- homotopy procedures are more verbose