Dev

dev@dynare.org

3402 discussions

Re: [DynareDev] Compiling Dynare under Windows
by Sébastien Villemot 03 Sep '10

03 Sep '10

On Fri, Sep 03, 2010 at 09:59:12AM +0100, George Perendia wrote: > This looks good and easier for a novice to use. > > 1) What about the old user requirements for windows (cygwin) not > listed any more such as: > * {{{gcc}}}, {{{gcc-g++}}} > * {{{gcc-mingw}}}, {{{gcc-mingw-g++}}} > * {{{gcc4}}}, {{{gcc4-g++}}} and {{{gcc4-gfortran}}} > * {{{readline}}} > * {{{libhdf5-devel}}} > > Are they not needed any more if one uses Mingw? No, they are not needed anymore for compiling Dynare with MinGW. But keeping them is not a problem either. > 2) What happened to Git users such as our development team? Do we > continue to use Git to download? I suppose we do but what then about > the precompiled libraries, will they be updateable via git too? Nothing has changed regarding Git. We still use it for the development of the Dynare source code. Concerning the precompiled binaries, I don’t plan to use a git repository for them, since they are going to change very rarely. Actually they’re going to change only if we introduce a new dependency for Dynare (for example when we added the GSL for the Markov-Switching code), or if we want to upgrade versions there. I will send an e-mail on the list if I update them. > 3) GitHowTo (still) suggests using "pull" to download but yuo say > you prefer us to use Fetch&Rebase. F&R may be easier to control from > the PM view but it does not merge local and remote changes as pull > does and so it seems that F&R overwrites any local changes if there > is a conflict which requires more work and control on our side or we > may loose our work. > > Could we use Pull and then Rebase instead F&R? Don’t use “pull” and then “rebase”, this will just mess up everything. The simple rule is the following : - if you have local commits which are not yet incorporated in the origin/master branch, then do a “fetch & rebase” - otherwise, if you have no local commit, a “pull” is fine (and in that case is equivalent to a “fetch & rebase”). Houtan is going to improve the wiki page on Git, I hope this will clarify things for you. Best, -- Sébastien Villemot CEPREMAP — http://www.cepremap.ens.fr Dynare project — http://www.dynare.org Phone: +33 1 40 77 49 90 PGP Key: 0xA6C029B9D06B2913D71C105EBE37E801FB6EFF8B (http://pgp.mit.edu/)

1 0

Fwd: RE: DSGE-net invoice
by Michel Juillard 02 Sep '10

02 Sep '10

FYI Best Michel ----- Forwarded message from giovanni.lombardo(a)ecb.int ----- Date: Wed, 1 Sep 2010 09:25:23 +0200 From: giovanni.lombardo(a)ecb.int Reply-To: giovanni.lombardo(a)ecb.int Subject: RE: DSGE-net invoice To: michel.juillard(a)ens.fr Dear Michel, thanks, I will forward it to the administration. By the way, the improvements in Dynare are impressive (and even more so those planned for 2010/2011). Well done! Best Gianni > -----Original Message----- > From: Michel Juillard [mailto:michel.juillard@ens.fr] > Sent: Tuesday 31 August 2010 21:54 > To: Lombardo, Giovanni > Subject: DSGE-net invoice > > > Dear Gianni, > > here is the invoice for DSGE-net > > All the best, > > Michel > Any e-mail message from the European Central Bank (ECB) is sent in good faith but shall neither be binding nor construed as constituting a commitment by the ECB except where provided for in a written agreement. This e-mail is intended only for the use of the recipient(s) named above. Any unauthorised disclosure, use or dissemination, either in whole or in part, is prohibited. If you have received this e-mail in error, please notify the sender immediately via e-mail and delete this e-mail from your system. ----- End forwarded message ----- --

3 2

Compiling Dynare under Windows
by Sébastien Villemot 01 Sep '10

01 Sep '10

Hi, I have updated the wiki page about the compilation of Dynare under Windows [1]. The new procedure is simpler, and has several improvements over the previous one : — The Markov-Switching (SWZ) MEX can now be built — Under Cygwin, the preprocessor binary no longer needs Cygwin DLL in the path — Possibility of compiling MEX files for 64-bit MATLAB with GCC and the build system (previously, it was necessary to have MS Visual C++) — Possibility of cross-compiling from Linux (not documented on the wiki). Hence, the snapshot is fully rebuilt (including MEX files) every night. Also note that the snapshot is now a Windows installer, similar to the one used for stable releases. Two problems remain, which I hope to fix soon : — Dynare++ binary is currently built with vanilla BLAS and LAPACK (instead of ATLAS) — Under MATLAB 64-bit, the mexErrMsgTxt() makes MATLAB crash. This is related to C++ exception propagation, which seems to have a problem in MinGW 64-bit. Hence if you call the MEX files without arguments, you will have a MATLAB crash. But running an error-free MOD-file works fine. Please let me know if you encounter any problem with the new procedure. Best, [1] http://www.dynare.org/DynareWiki/BuildingDynareFromSource -- Sébastien Villemot CEPREMAP — http://www.cepremap.ens.fr Dynare project — http://www.dynare.org Phone: +33 1 40 77 49 90 PGP Key: 0xA6C029B9D06B2913D71C105EBE37E801FB6EFF8B (http://pgp.mit.edu/)

1 0

[Trac] #118: Build system for SWZ DLL fails under Cygwin
by Trac Dynare 26 Aug '10

26 Aug '10

#118: Build system for SWZ DLL fails under Cygwin --------------------------------------+------------------------------------- Reporter: sebastien | Owner: sebastien Type: bug | Status: new Priority: major | Milestone: 4.2 Component: Build system / Packaging | Version: Keywords: | --------------------------------------+------------------------------------- The problem comes from the GSL: the package shipped with Cygwin does not link with MATLAB DLLs, which are compiled with MinGW even under Cygwin. The solution is probably to use a GSL compiled for MinGW: http://gnuwin32.sourceforge.net/packages/gsl.htm -- Ticket URL: <https://www.dynare.org/trac/ticket/118> Dynare <http://www.dynare.org> The Dynare project

1 1

[Trac] #96: Problem with struct2local.m
by Trac Dynare 25 Aug '10

25 Aug '10

#96: Problem with struct2local.m --------------------------+------------------------------------------------- Reporter: sebastien | Owner: Type: bug | Status: new Priority: critical | Milestone: Component: Core M-files | Version: 4.1.1 Keywords: | --------------------------+------------------------------------------------- The function struct2local.m doesn't work as expected when a function exists in the path with the same name than a variable to be unpacked. After the call to struct2local, MATLAB gives precedence to the function over the local variable. This behavior is apparently in contradiction with the precedence rules described on: http://www.mathworks.com/access/helpdesk/help/techdoc/matlab_prog/f10-60956… However it is present on most versions of MATLAB, and should be taken as a normal behavior of MATLAB. A possible fix is to drop the use of struct2local, and to unpack structures by hand. -- Ticket URL: <https://www.dynare.org/trac/ticket/96> Dynare <http://www.dynare.org> The Dynare project

1 5

[Trac] #124: write additional documentation for 3rd order
by Trac Dynare 24 Aug '10

24 Aug '10

#124: write additional documentation for 3rd order ---------------------------+------------------------------------------------ Reporter: michel | Owner: Type: bug | Status: new Priority: major | Milestone: 4.2 Component: Documentation | Version: Keywords: | ---------------------------+------------------------------------------------ - mention use_dll in order option of stoch_simul in Reference Manual - document the output variables of 3rd order (on the Wiki?) -- Ticket URL: <https://www.dynare.org/trac/ticket/124> Dynare <http://www.dynare.org> The Dynare project

1 1

[Trac] #123: add external_function to the Reference Manual
by Trac Dynare 26 Jul '10

26 Jul '10

#123: add external_function to the Reference Manual ---------------------------+------------------------------------------------ Reporter: michel | Owner: Type: enhancement | Status: new Priority: major | Milestone: 4.2 Component: Documentation | Version: Keywords: | ---------------------------+------------------------------------------------ -- Ticket URL: <https://www.dynare.org/trac/ticket/123> Dynare <http://www.dynare.org> The Dynare project

1 1

Re: [DynareDev] Dynare Octave problem
by George Perendia 21 Jul '10

21 Jul '10

Update FYI: After all, it seems like I can make octave debug prompt appear and work if I set dbstop (or keyboard) and then run e.g. octave> sweuromodel (i..e as..m file) instead (i.e. and after) octave> dynare sweuromodel.mod noclearall cygwin it looks like running dynare "upsets" switch to octave debug mode (i.e. same problem occur when running via dynare even if I do not use use_dll option) Best George On 21/07/2010 09:47, George Perendia wrote: > HI Sebastien > > Thanks for all your answers > > Problem is however that each described method of entering debug mode > results in my prompt and cygwin window freezing and the only thing I > can do is C-c to exit the Dynare application. I can however make the > dll print matrices to cout under octave and I can put pause and other > suitable commands in .m files to pause or control the execution so > that helps but slows down my debugging. > > Best > George > > On 20/07/2010 21:17, Sébastien Villemot wrote: >> Hi, >> >> On Mon, Jul 19, 2010 at 08:11:04PM +0100, George Perendia wrote: >>> Another issue is that when octave hits dbstop: the prompt and >>> keyboard freeze at the first line of the file dbstop is set in >>> without promptimg "debug>" mode (and not at the line I specified >>> using dbstop<fname>,<lineno>.: >> I don’t have much experience with debugging under Octave, and I don’t >> know if >> it is supposed to behave like MATLAB. I can only point you to the >> part of >> Octave manual which deals with debugging : >> >> http://www.gnu.org/software/octave/doc/interpreter/Debugging.html >> >> Best, >> > > --

1 0

[Trac] #108: Create an interface for DSGE-Var
by Trac Dynare 06 Jul '10

06 Jul '10

#108: Create an interface for DSGE-Var --------------------------+------------------------------------------------- Reporter: sebastien | Owner: sebastien Type: enhancement | Status: new Priority: major | Milestone: 4.2 Component: Preprocessor | Version: Keywords: | --------------------------+------------------------------------------------- Needs an option of estimation, which takes a value: - if the value is numeric, means that the weight of the DSGE prior is calibrated to that value - if the value is a string, means that this weight is estimated as this given parameter Also needs an option for number of lags. See: http://www.dynare.org/DynareWiki/DsgeVar -- Ticket URL: <https://www.dynare.org/trac/ticket/108> Dynare <http://www.dynare.org> The Dynare project

1 5

Re: [DynareDev] [Trac] #24: Add Markov-Switching
by Trac Dynare 01 Jul '10

01 Jul '10

#24: Add Markov-Switching ---------------------------+------------------------------------------------ Reporter: sebastien | Owner: houtanb Type: feature | Status: accepted Priority: major | Milestone: 4.2 Component: Core M-files | Version: Resolution: | Keywords: ---------------------------+------------------------------------------------ Comment(by houtanb): Replying to [comment:1 sebastien]: > Note: the SWZ build system currently works on Linux and MacOS, but not Cygwin. The problem has to do with the GSL whichs creates conflicts with MATLAB libraries. > Ported to mex and made to work on Mac and Linux in commits between ac0f324d6fc73e281d46775c231f9bc0f21dc99e and 359b90d4fcd9a08ddc12524724b76d0fa6e202c6 -- Ticket URL: <https://www.dynare.org/trac/ticket/24#comment:3> Dynare <http://www.dynare.org> The Dynare project

1 0

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

Dev