Navigation:  Menu Commands > File Menu >

 

Parallel Computing Toolbox

 

Print this Topic Previous pageReturn to chapter overviewNext page

If you have matlab's Parallel computing toolbox installed, you can open or close a parallel pool from the file menu. YADA always makes use of your default profile for the pool.

When you have opened a parallel pool, YADA makes use of the parfor function when running certain loops provided that it can use at least 2 computational threads (number of workers). The tools currently supported for this are listed below:

The mutation step of the SMC with likelihood or data tempering posterior sampler for DSGE and DSGE-VAR models, as well as the correction step of the SMC with data tempering posterior sampler.
The computation of importance weights and t-densities when using importance sampling based on the MitISEM algorithm.
Impulse response functions for DSGE models.
Impulse response functions for DSGE-VAR models.
Conditional variance decompositions for DSGE models.
Variance decompositions for DSGE models.
Variance decompositions for DSGE-VAR models.
CRPS and ES computations for DSGE models under rational expectations and subject to adaptive learning.
Unconditional predictive distributions for DSGE models under rational expectations and subject to adaptive learning.
Conditional predictive distributions for DSGE models under rational expectations.
Zero (effective) lower bound computations.
Inverse Hessian with finite differences.
Predictive likelihood estimation and marginal predictive moments estimation with prior and posterior draws for DSGE-VAR models, and for DSGE models under rational expectations and adaptive learning.

 

 


Page url: https://www.texlips.net/yada/html/index.html?parallel_computing_toolbox.htm