state of dealii in complex arithmetics

[Juan Carlos Araujo Cabarcas]

Dear all,

There are features in SLEPc 3.7 that I am interested in, and will try to make a fresh re-installation soon.

I have been using deal.II 8.3 with complex arithmetic’s from the branch:
    git checkout branch_petscscalar_complex
and I wonder about the state of that branch, or if it has been merged to the master branch.

Thanks in advance,

Juan Carlos Araújo,
Umeå Universitet

[Wolfgang Bangerth]

Juan Carlos,

> There are features in SLEPc 3.7 that I am interested in, and will try to
> make a fresh re-installation soon.
>
> I have been using deal.II 8.3 with complex arithmetic’s from the branch:
> git checkout branch_petscscalar_complex
> and I wonder about the state of that branch, or if it has been merged to
> the master branch.

I think the majority was merged. Give it a shot and see what breaks and
doesn't work yet, and we'll see how that can be fixed. No better way to
fixing bugs than trying to actually use things.

Best
W.


--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/

[Denis Davydov]

Hi Juan Carlos,

Have a look at this issue https://github.com/dealii/dealii/issues/2033

you will see some "open" issues linked to it, which are known limitations when using complex arithmetics.

The main one has to do with constraints. Currently we only allow real-valued constraints (Dirichlet BC + hanging nodes).

Regards,

Denis.

[Denis Davydov]

On Tuesday, May 23, 2017 at 5:22:44 PM UTC+2, Juan Carlos Araujo Cabarcas wrote:

Dear all,

There are features in SLEPc 3.7 that I am interested in, and will try to make a fresh re-installation soon.

I have been using deal.II 8.3 with complex arithmetic’s from the branch:
    git checkout branch_petscscalar_complex
and I wonder about the state of that branch, or if it has been merged to the master branch.

yes, everything in that branch is in the main repo from 8.4.2 or even earlier.

[Juan Carlos Araujo Cabarcas]

Thanks, Denis and Wolfgang for the quick reply.
I will give it a try and report back whatever issue I encounter.