Clusters of free or limited access for big models

Estimation
1

My model has 283 equations and 23 observed variables, so the estimation is very slow. Takes something of a day. And also, I want to compute root-mean-square error for 30 windows.

I would like to know if there is some clusters of free or limited acces of institutions to run dynare faster from my laptop, something like remote access?

Thanks for your help.

2

What computation takes a day? Finding the posterior mode? Running the Metropolis iterations? How many iterations

3

Thanks for the answer, dear Michel Juliard,

Finding the posterior mode takes something 20 hours and the Metropolis takes nearly five hours. I have 74 parameters.

These are my options:

estimation(
datafile=‘consolidado’,
presample=0,
mode_compute = 5, optim = (‘Hessian’,2),
mode_check,
diffuse_filter,
mh_replic=50000,
mh_drop=0.333,
//mode_file = mymode,
mh_nblocks=1,
posterior_sampling_method = ‘random_walk_metropolis_hastings’,
tex
)