Dev July 2011

dev@dynare.org

6 participants
72 discussions

[Trac] #127: Output 2nd derivatives of static model when doing identification
by Trac Dynare 19 Aug '11

19 Aug '11

#127: Output 2nd derivatives of static model when doing identification --------------------------+------------------------------------------------- Reporter: sebastien | Owner: sebastien Type: feature | Status: new Priority: major | Milestone: 4.3 Component: Preprocessor | Version: Keywords: | --------------------------+------------------------------------------------- This is necessary for global identification. The code for 2nd static derivatives already exists (used for planner_objective). -- Ticket URL: <https://www.dynare.org/trac/ticket/127> Dynare <http://www.dynare.org> The Dynare project

1 2

[Trac] #183: mex file for the static model
by Trac Dynare 19 Aug '11

19 Aug '11

#183: mex file for the static model --------------------------+------------------------------------------------- Reporter: stepan | Owner: houtanb Type: enhancement | Status: new Priority: major | Milestone: 4.3 Component: Preprocessor | Version: Keywords: | --------------------------+------------------------------------------------- Dynare does not generate a mex file for the static model when option is use_dll is set. It would be nice to have this file because a lot of time is spent here when a closed form solution is not specified for the steady state. -- Ticket URL: <https://www.dynare.org/trac/ticket/183> Dynare <http://www.dynare.org> The Dynare project

1 2

[Trac] #175: bug in discretionary_policy_engine.m
by Trac Dynare 19 Aug '11

19 Aug '11

#175: bug in discretionary_policy_engine.m --------------------------+------------------------------------------------- Reporter: michel | Owner: michel Type: bug | Status: new Priority: major | Milestone: Component: Core M-files | Version: Keywords: | --------------------------+------------------------------------------------- - first function is not terminated by 'end' contrarily to the other ones. Matlab (r2010b) is missing it. - W may not be initialized correctly. ./tests/discretionary_policy/dennis_1.mod is an example. Because of above, Matlab is missing it, but Octave detects it. -- Ticket URL: <https://www.dynare.org/trac/ticket/175> Dynare <http://www.dynare.org> The Dynare project

1 1

Re: [DynareDev] EAGLE in Dynare 4.2.1
by Sébastien Villemot 22 Jul '11

22 Jul '11

Dear Matija, Lozej Matija <Matija.Lozej(a)bsi.si> writes: > Is the ECB's EAGLE model compatible with Dynare version 4.2.1.? When I > try to compute the steady state (dynare steady1.mod), I get the > message (after some claculations are done) "Fatal error in bytecode". I know there are some issues with Eagle and recent versions of Dynare. I am not completely aware of them, so I forward this email to Ferhat Mihoubi who is in charge of the relevant code and will give you a more precise answer. In the meanwhile, I suggest that you give a try to the unstable version of Dynare, that can be downloaded in the form of a daily snapshot on the website. Maybe the bug that you encountered is fixed in the unstable version. Best, -- Sébastien Villemot CEPREMAP - Paris School of Economics Homepage: http://www.dynare.org/sebastien Landline phone: +33 1 40 77 49 90 SIP phone: sebastien.villemot(a)ekiga.net PGP Key: 2069 1DFC C2C9 8C47 9529 84EE 0001 8C22 381A 7594

2 1

Re: [DynareDev] histval et stoch_simul
by Stéphane Adjemian 18 Jul '11

18 Jul '11

Il n'y a pas d'exemple (j'ai essayé avec example1.mod)... Le problème est évident car dans le code on passe systématique dr.ys comme condition initiale. J'imagine que dans ce cas il suffit de passer la première colonne de oo_.endo_simul. Mais il faudrait un flag pour signaler que le bloc histval est employé et il faudrait que le préprocesseur ne sorte pas en crashant... Je réponds donc sur le forum qu'il s'agit d'un bug de dynare... Pour le second point on verra plus tard. J'ai commencé à écrire une classe time series compatible avec octave... Il faut prévoir quelque chose pour ça. à+ Stéphane. Le 18/07/11 15:48, Michel Juillard a écrit : > Salut Stéphane; > > On 7/18/2011 3:20 PM, Stéphane Adjemian wrote: >> Salut Michel, >> >> Une petite question liée au forum... >> >> Si on veut démarrer une simulation stochastique avec une condition >> initiale, je pensais qu'il était possible d'utiliser le bloc histval... >> Mais ça ne marche pas (on se fait jeter par le préprocesseur). En >> regardant le code, rien n'est prévu pour ça, et on ne peut initialier >> qu'avec l'état stationnaire déterministe... Connais tu la raison de >> cette restriction ? Cela ne me semble pas tellement en ligne avec la >> doc... >> > Il s'agit d'une erreur. On devrait pouvoir utiliser histval dans ce but. > Peux-tu m'envoyer l'exemple? > > >> Sinon, je ne comprends pas pourquoi on vire la condition initiale des >> simulations stochastiques. Cela n'aide pas tellement à comprendre ce qui >> se passe (j'aime bien voir le niveau initial des états). >> > > C'est pour créer des matrices (vecteurs) qui la taille de l'option > 'periods'. Je crois que la seule bonne solution serait de retourner les > résultats avec des fonctions et de gérer le stockage de manière plus > rationnelle > > amicalement > > Michel > >> à+ Stéphane. >> > -- Stéphane Adjemian Université du Maine, GAINS & CEPREMAP http://www.dynare.org/stepan TEL (Cepremap) +33(0)1 40 77 84 19 TEL (Gains) +33(0)2 43 83 31 35

2 1

Fwd: [Trac] #194: Clarify Slicot licensing terms
by Stéphane Adjemian 16 Jul '11

16 Jul '11

Hi Sébastien, Slicot library is still not used in the estimation dll... I downloaded succesfully the version 5 of Slicot this morning (currently Dynare distributes version 4). On Slicot's web site the library is still presented under a Gnu GPL licence. I will contact the authors next week to see if they plan to change the licence... Yesterday you told me that it was not possible to change a GPL licence, because anyone can fork... So I guess that it is not an issue if we distribute old versions of Slicot, even if they choose another licence for the next versions (>5). Best, Stéphane. -------- Message original -------- Sujet: [DynareDev] [Trac] #194: Clarify Slicot licensing terms Date : Fri, 15 Jul 2011 20:25:54 -0000 De : Trac Dynare <root(a)dynare.org> Répondre à : List for Dynare developers <dev(a)dynare.org> Copie à : dev(a)dynare.org #194: Clarify Slicot licensing terms ----------------------------+----------------------------------------------- Reporter: sebastien | Owner: george Type: bug | Status: new Priority: critical | Milestone: Component: Estimation DLL | Version: Keywords: | ----------------------------+----------------------------------------------- Currently Dynare (unstable) distributes its own copy of libslicot (www.slicot.org). It is used in the Kalman Steady State DLL (at least). As of July 15, 2011, it is no longer possible to download libslicot from its website. It says that the library is currently undergoing a relicensing. We need to clarify this with Slicot authors to see if we can continue distributing it with Dynare. -- Ticket URL: <https://www.dynare.org/trac/ticket/194> Dynare <http://www.dynare.org> The Dynare project _______________________________________________ Dev mailing list Dev(a)dynare.org https://www.dynare.org/cgi-bin/mailman/listinfo/dev -- Stéphane Adjemian Université du Maine, GAINS & CEPREMAP http://www.dynare.org/stepan TEL (Cepremap) +33(0)1 40 77 84 19 TEL (Gains) +33(0)2 43 83 31 35

3 3

An update on parallelization on GPU
by Sébastien Villemot 15 Jul '11

15 Jul '11

Hi, This morning I have documented myself a little bit on the subject of exploiting parallelization on GPUs (Graphical Processing Units). George had made a presentation on this topic a few years ago, and things have evolved a little bit since. There seem to be now two major competing programming frameworks for parallelization on GPUs (and possibly CPUs also): - CUDA, developed by Nvidia [1] - OpenCL, presented by a consortium (including Apple, AMD, ATI, Nvidia) [2] CUDA was initially only available on Nvidia graphics cards (see the packages "nvidia-cuda-*" in Debian, which contain the non-free implementation of CUDA for Nvidia cards). There is now a free implementation of CUDA, called GPU Ocelot [3], which looks promising. It is able to run CUDA programs on an heterogenous cluster of GPUs (Nvidia, AMD, Intel) and CPUs (multi-core x86). It is not yet packaged in Debian, but it seems very interesting and we should give it a try. On the long term, OpenCL looks more promising, but there is currently no free implementation of it. There is a Debian package called "opencl-headers", but it contains only the headers, since the binary implementation is non-free. There are also Debian packages containing the Nvidia non-free implementation of OpenCL (see "nvidia-opencl-*"). I am not aware of any ongoing project to build a free implementation of OpenCL. To summarize: whoever has the interest and time in our team should give a try to GPU Ocelot to see if we can take advantage of it in Dynare. Don't hesitate to react to this mail if you know things on this topic of which I am not aware. Best, [1] http://en.wikipedia.org/wiki/CUDA [2] http://en.wikipedia.org/wiki/OpenCL [3] http://code.google.com/p/gpuocelot/ -- Sébastien Villemot CEPREMAP - Paris School of Economics Homepage: http://www.dynare.org/sebastien Landline phone: +33 1 40 77 49 90 SIP phone: sebastien.villemot(a)ekiga.net PGP Key: 2069 1DFC C2C9 8C47 9529 84EE 0001 8C22 381A 7594

2 1

Fwd: Re: Suggestion
by Michel Juillard 14 Jul '11

14 Jul '11

... Thanks Johannes. This is an excellent suggestion. I pushed the modification to GIT. All the best Michel On 07/14/2011 12:01 PM, Johannes Pfeifer wrote: > /Please check i) if all parameters occurring in these equations are > defined and ii) that no division by an endogenous variable initialized > to 0 occurs. /

1 0

[Trac] #189: identification options
by Trac Dynare 12 Jul '11

12 Jul '11

#189: identification options --------------------------+------------------------------------------------- Reporter: houtanb | Owner: houtanb Type: bug | Status: new Priority: major | Milestone: 4.3 Component: Preprocessor | Version: Keywords: | --------------------------+------------------------------------------------- Add lik_init and/or diffuse_filter -- Ticket URL: <https://www.dynare.org/trac/ticket/189> Dynare <http://www.dynare.org> The Dynare project

1 3

[Trac] #187: macrovariable passed on the command line (-D option)
by Trac Dynare 12 Jul '11

12 Jul '11

#187: macrovariable passed on the command line (-D option) --------------------------+------------------------------------------------- Reporter: michel | Owner: sebastien Type: bug | Status: new Priority: major | Milestone: 4.3 Component: Preprocessor | Version: Keywords: | --------------------------+------------------------------------------------- It works for numerical values but not for strings. In the test.mod file: @#if var1 == "a" @#echo "OK" @#else @#echo "NO" @#endif Calling dynare test -Dvar1=a will fail with with unknown variable a (without any Dynare instruction, it is normal that test.mod fails at the end, but with an error message about unexpected $end) -- Ticket URL: <https://www.dynare.org/trac/ticket/187> Dynare <http://www.dynare.org> The Dynare project

1 4

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

Dev July 2011