Solve Poisson Problem with Finite Difference using parallelization

3 vues (au cours des 30 derniers jours)
Kilian Thomas
Kilian Thomas le 8 Août 2014
Commenté : Kilian Thomas le 8 Août 2014
Hi, I am trying to solve the poisson problem with finite difference. However, depending on the grid size, the calculation may take time. That's why I want to using parallelism. I already solved this problem using and I would like to do it on matlab. Is there any way in matlab to solve two subproblems in parallel that need to communicate with each other?
Thank you

Réponse acceptée

Edric Ellis
Edric Ellis le 8 Août 2014
You might consider using SPMD blocks within MATLAB - these run blocks of code in parallel, and allow communication using the labSend / labReceive family of functions.
  1 commentaire
Kilian Thomas
Kilian Thomas le 8 Août 2014
thank you. I think it could work with this. I only need to have the parallel tool now.
thank you :)

Connectez-vous pour commenter.

Plus de réponses (1)

Ahmet Cecen
Ahmet Cecen le 8 Août 2014
Poisson equation boils down to a simple Ax=b problem. If you have the parallel computing toolbox, using mldivide (or the plain \ operator) will automatically exploit all available threads. Otherwise, I would look into 3rd party MP/MPI implementations for MATLAB.
  3 commentaires
Ahmet Cecen
Ahmet Cecen le 8 Août 2014
Gauss Seidel is an iterative method to solve LINEAR problems of the exact form Ax=b. Your problem is very much linear. The choice of not using mldivide is I am guessing due to the fact that you are interested in a transient solution rather than steady state, which was not clear in your question.
Kilian Thomas
Kilian Thomas le 8 Août 2014
Yes sorry. You are totally right. I will have a look at your solution too. It could be much faster. Thank you :)

Connectez-vous pour commenter.

Catégories

En savoir plus sur Programming dans Help Center et File Exchange

Produits

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by