Christophe Prud'homme

icon indicating a website link http://www.cemosis.fr icon indicating an email [email protected]

icon company University of Strasbourg icon location Strasbourg icon location Professor in applied mathematics at University of Strasbourg. Director of the Cemosis platform. Leader of NumPEx/Exa-MA Methods and Algorithms for Exascale

Results 221 comments of Christophe Prud'homme

Clean Code

@vincentchabannes the configuration we did was not taken into account (wrong directory), now it is good and type:refactoring is one of the label that pin an issue

Expression of second derivative (by component) with vectorial field

I don't think `dxx ...` is a good idea either. best is to have a function like you propose ```c++ d2(u,i,j) ```

Add code to solve heat equation as a sandbox for radiative heat transfer

@MichelFouquembergh @alixkien the 2D/3D code is there. There remains to finish implementing the RHT boundary conditions

support enabling/disabling writing postprocess on disk

@vincentchabannes if we compute extra fields in the postprocess step, are they available from the API ? We may want to compute the fields, use them in some other computations...

support enabling/disabling writing postprocess on disk

> I am not sure to understand. Field is a finite element function? If i run a time step of the fluid toolbox, and have an expression computing say the...

Support better exception handling and error messages

@vincentchabannes I move this to v107.

Replace Boost::MPL

I have already compared and mp11 is the right choice imo I already played with it in a branch, I can push eventually @metivett boost::hana is not dropped, boost::hana is...

Replace Boost::MPL

boost::hana is not pure tmp, it crosses the boundary. in many cases we just need pure tmp, but I agree that we use boost::hana more often. We still need to...

Replace Boost::MPL

I would go for mp11 which is maintained by P Dimov, it is well documented, and I already used it in a few places. range-v3 is there until we get...

add support for random partitioning

@vincentchabannes you have setup something like that no ? what is the process to stress the partitioner and feelpp mpi implementation