Post Processor for a mutli-component system

[SALMAN]

Dear all

I am using dealii to solve a multi component system has its own fe, dofHandler, and solution vector. I am using a matrix free implementation for the entire program.

I am using the Postprocessor class to derive quantities that depend on all of the components.

The Postprocessor works fine in serial implementation. However, when trying to run using MPI the output is producing wrong results.

I created a procedure to generate a joint solution vector and a joint handler following tutorial 32. However, this tutorial uses Trillinos vectors and I am using dealii’s own distributed vector and I am suspecting that this might be problem.

Hence, my question is how to properly generate a joint solution vector while using dealii's distributed::vector class.

I have attached a text file that contains the procedure that was used to generate a the joint vector and I would appreciate if anyone would point out the mistakes that I made

Thank you

​

[Martin Kronbichler]

I did not check all code in detail, but what I observe is that you write into the joint_solution vector without ghosts. Now this would likely result in a run time error (for countinuous elements and in debug mode), rather than wrong output. But either way, I would recommend you move the "locally_relevant_joint_solution" and the way you initialize it to the place where you declare "joint_solution" now, fill the data, and instead of calling "compress" you simply call "update_ghost_values". 

If that does not help, could you please specify in more detail where the output is wrong: Is it just cells near processor interfaces? What kinds of elements are you using and what kind of information is the postprocessor computing?

Best,
Martin

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dealii+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/c4beae48-c449-4b03-85df-1406bebc11dbn%40googlegroups.com.

[SALMAN ALRAKAN]

Prof. Kronbichler

First, I would like to thank you for the development of matrix free library.

Sir, I am extremely thankful for your response. I have modified the procedure and it's working perfectly.

I am simulating the evolution of Zr-delatH under mechanical loads by developing a phase field model.

I am solving for four main fields, temperature, displacement, concentration, and in addition a non-conserved order parameter.

I needed a post processor to evaluate the field's solutions and the  chemical , interfacial, elastic, energies along with their gradients.

Due to my short experience in dealii vector classes, I suspected that this was the problem and I am extremely thankful to you.

Thank you

You received this message because you are subscribed to a topic in the Google Groups "deal.II User Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dealii/ED8T-J3qyu4/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dealii+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/71317c8d-a236-ccf5-3b03-0ebb3a1d95da%40gmail.com.