Kratos: coordinate_transformation_utilities.h

[jrubiogonzalez]

I'm working with the BDF scheme, and it is using the coordinate_transfromation_utilities.h. I see it is using the old flag IS_STRUCTURE varible. Is it worth it to adapt it to to use the STRUCTURE flag? What about the NORMAL variable used? Is it worth moving it to the Non-Historical DD.BB?

Need your comments: @jcotela, @rubenzorrilla @RiccardoRossi @KratosMultiphysics/fluid-dynamics

[jcotela]

I'm completely for moving from IS_STRUCTURE to the STRUCTURE flag, this has been long in planning but we are always reluctant because it is a quite wide change. Do you think it would be worth it to start the changes with your (@jrubiogonzalez's) BDF and then propagating the changes to the other solvers?

Regarding the normals, I'm a bit more ambivalent. I would like to have a better idea of the tradeoffs before commiting to a change (faster access from historical DB vs less memory in non-historical DB). Does anyone have any numbers on which to base an informed decision?

[jrubiogonzalez]

I volunteer as a guniea pig for the being the first in using the changes. But maybe we can plan something to provide a smooth transition for everyone.

It would be good to have the information about the trade-off in Bs's. ( @pooyan-dadvand ?)

[maceligueta]

+1 for converting the variable IS_STRUCTURE into the STRUCTURE flag

[rubenzorrilla]

As @maceligueta points, +1 for getting rid of the IS_STRUCTURE var. @jrubiogonzalez, which BDF scheme are you using? It is a serial or an MPI one? . This has nothing to do with the IS_STRUCTURE and NORMAL discussion, but yesterday Jordi and me discovered a strange behaviour of the trilinos_gear_scheme used in combination with my monolithic trilinos solver...

[jrubiogonzalez]

regular residualbased_predictorcorrector_velocity_bdf_scheme_turbulent.h

[pooyan-dadvand]

👍 for STRUCTURE flag

About the normal I think @RiccardoRossi had a use case where the old normal in previous time step was also needed. Am I right Riccardo?

[RiccardoRossi]

well, if you want to use a Crank Nicholson scheme and to be very exact, than you would need the normal at n+1/2 which would require knowing both the current normal and the old one.

[loumalouomega]

@jrubiogonzalez Did you solved the problem?

[RiccardoRossi]

nope it is active

On Wed, May 31, 2017 at 3:41 PM, Vicente Mataix Ferrándiz <
[email protected]> wrote:

@jrubiogonzalez https://github.com/jrubiogonzalez Did you solved the
problem?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/KratosMultiphysics/Kratos/issues/214#issuecomment-305189873,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AHr7ESD2QKRsdb6iZUCEU6em1ZXFQKPBks5r_W34gaJpZM4MrRm6
.

--

Riccardo Rossi

PhD, Civil Engineer

member of the Kratos Team: www.cimne.com/kratos

Tenure Track Lecturer at Universitat Politècnica de Catalunya,
BarcelonaTech (UPC)

Full Research Professor at International Center for Numerical Methods in
Engineering (CIMNE)

C/ Gran Capità, s/n, Campus Nord UPC, Ed. B0, Despatx 102

(please deliver post to the porters of building C1)

08034 – Barcelona – Spain – www.cimne.com -

T.([phone] skype: rougered4

http://www.cimne.com/

https://www.facebook.com/cimne http://blog.cimne.com/
http://vimeo.com/cimne http://www.youtube.com/user/CIMNEvideos
http://www.linkedin.com/company/cimne https://twitter.com/cimne

Les dades personals contingudes en aquest missatge són tractades amb la
finalitat de mantenir el contacte professional entre CIMNE i voste. Podra
exercir els drets d'accés, rectificació, cancel·lació i oposició,
dirigint-se a [email protected]. La utilització de la seva adreça de
correu electronic per part de CIMNE queda subjecte a les disposicions de la
Llei 34/2002, de Serveis de la Societat de la Informació i el Comerç
Electronic.

Imprimiu aquest missatge, només si és estrictament necessari.
http://www.cimne.com/

[RiccardoRossi]

@KratosMultiphysics/technical-committee i think we should push this for release 6.0

[RiccardoRossi]

@jcotela @rubenzorrilla @jrubiogonzalez do we make a try to solve this?

[pooyan-dadvand]

Is there any move in changing the IS_STRUCURE to flag? It is in release milestone.

[jcotela]

@pooyan-dadvand not really, but I think we have the time to make it if we decide it is a priority. I don't expect it to be a hard change, but we all need to agree and change it at once (the big issue is that the change will have to be ported to c2c too, otherwise we would break it).

[jcotela]

@KratosMultiphysics/technical-committee In trying to implement this change, I have realized that we need #1277 first for the flag-based version to work cleanly in MPI. I would therefore suggest giving priority to that feature before changing this. I do not know if we can make it in time for release 6, however.

[loumalouomega]

Is this solved?

[jcotela]

@loumalouomega nope, as it says in the thread, this is blocked until we implement #1277.

[philbucher]

the syncronization is available now, I think you can go forward with this now

[jcotela]

@jrubiogonzalez @KratosMultiphysics/altair This issue can be solved now. Is it still a priority for you? I can go ahead and do the changes in the core and FluidDynamics but you'll need to follow the changes in your applications.

[jrubiogonzalez]

No problem, we will follow. Thanks.

[RiccardoRossi]

guys, i do think that this is important, nevertheless this will affect the implementation that is currently been done by @roigcarlo so i would like his comment on this

[roigcarlo]

@RiccardoRossi The code I have is using the STRUCTURE flag already. It does not have any mpi "interaction" so go ahead if you want.

[RiccardoRossi]

perfect!

[jcotela]

One question: for wall conditions (and the related coordinate rotation) it may make sense to use SLIP instead of STRUCTURE. Should I go ahead and use it? We can still assign both flags in the corresponding setup process if desired.

[RiccardoRossi]

+1 for me calling it SLIP instead of STRUCTURE

[jrubiogonzalez]

In fact we are using both, I think it works for us but we should check.

[rubenzorrilla]

+1 for SLIP