- Implementation of the simulation of the block decomposed model using sparse matrix using MATLAB code:
New option in model : sparse
Example:
Model(SPARSE);
….
End;
- Implementation of the simulation of the block decomposed model using sparse matrix for static model.
- Krylov methods are implemented to solve the linear model J.y=b, with J the Jacobian matrix at each step of the Newton Method. Two Krylov's solvers are considered :
+ Generalized Minimum Residual Method (GMRES)
+ BiConjugate Gradients Stabilized (BICGSTAB)
To call them the SIMULATE command has to be completed with a new option : METHOD which could have three different values :
+ LU : (default value) Sparse LU resolution
+ GMRES
+ BICGSTAB
Example:
simul(periods=100, method=lu);
or simul(method=Gmres);
- Correction of various bugs in “simulate.dll”.
- Correction of simk.m
- Correction allowing a list of VALUES or PERIODS not necessarily separated by a comma.
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1447 ac1d8469-bf42-47a9-8791-bf33cf982152
* in the .m file, corrected "if exist" and "mex" lines
* in the .c files, fixed problem with model local variables
* reverted Fehrat's change: power operator in C is pow(), not pow1()
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1301 ac1d8469-bf42-47a9-8791-bf33cf982152
* implemented handling of mod file local variables (eModFileLocalVariable)
* renamed "local parameters" to model local variables (eModelLocalVariables)
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1253 ac1d8469-bf42-47a9-8791-bf33cf982152
* removed Expression class; the "ExprNode" class is therefore now used everywhere
* removed interprete class, and replaced it by a method in ExprNode and an global evaluation context in ModFile
* fixed breakage of SparseDLL / Block decomposition code introduced in previous revision
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1204 ac1d8469-bf42-47a9-8791-bf33cf982152
* reorganised code in ModelTree so that block decomposition and SparseDLL code is clearly separated
* replaced DataTree::offset by ModelTree::mode, using an enumeration of the three modes which is more explicit
* reorganised ExprNode::writeOutput method by using a sixfold enumeration type (ExprNodeOutputType) corresponding to the 6 different contexts in which an expression can be written
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1203 ac1d8469-bf42-47a9-8791-bf33cf982152
* added third order derivatives in dynamic file (triggered by order=3 option of stoch_simul/olr/osr/ramsey_policy)
* don't output hessian in dynamic file if order=1 is specified
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1180 ac1d8469-bf42-47a9-8791-bf33cf982152
* reshaped DataTree/ModelTree using a more object-oriented approach
* improved expression sharing algorithm (temporary terms in output files), resulting in smaller output files for big models
* optimized performance (both in computing time and memory consumption)
* bugfix: the expressions stored in local parameters of models (pound expressions) are now derived
* bugfix: cosh(0) = 1
* removed some useless parentheses in output
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1148 ac1d8469-bf42-47a9-8791-bf33cf982152
* added planner_objective statement, which triggers the creation of filename_objective_static.m
* added computing pass
* added the possibility of writing the hessian in static output file
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1126 ac1d8469-bf42-47a9-8791-bf33cf982152
* added an embryonic checking pass after parsing
* modified the decision of which derivatives to compute:
- if a simul statement is present, compute the jacobian vs endogenous variables only, and no hessian
- if a stoch_simul/estimation/olr/osr statement is present, compute the jacobian vs all variables, plus the hessian
- if the two previous conditions are simultaneously fulfilled, abort with an error
- if none of these two conditions are fulfilled, abort with an error (nothing to compute)
* fixed a minor bug with the size of the jacobian matrix when computing it only vs endogenous variables
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1120 ac1d8469-bf42-47a9-8791-bf33cf982152
* cleaned and simplified the code
* incorporated VariableTable as a private member of DataTree/ModelTree
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1118 ac1d8469-bf42-47a9-8791-bf33cf982152
* clearly separated the parsing process from the output writing, using an abstract intermediary representation of the mod file
* created the Statement class, and its derivatives, for that purpose
* fixed bug with "corr" statements in "shocks" blocks (variables indexes given to matlab were incorrect)
* fixed bug with "optim" option of "estimation" (this option was not passed to matlab)
* fixed bug with "dynatype" and "dynasave" (variable list was not passed correctly to matlab)
* fixed bug with "calib" and "calib_var" (they were unusable because of flex misspecification)
* fixed segmentation fault with "estimated_params_bound" and "estimation_params_init"
git-svn-id: https://www.dynare.org/svn/dynare/dynare_v4@1117 ac1d8469-bf42-47a9-8791-bf33cf982152
ferhat
sebastien
michel