Jørgen Schartum Dokken

It is not yet in a released version of Paraview (as far as Im aware). However, i agree that once there is a binary with the support we should implement...

> My compiled version of Paraview is a binary that supports it. > > > > P.s. :stuck_out_tongue: What version is that? I looked at the source code the other...

> The problem is that these functions are not generally supported in the C math library, also often not implemented in C++, either. We can add a call to something...

Then the question is, should they exist in ufl? firedrake doesn’t seem to support them https://github.com/firedrakeproject/tsfc/blob/90c20c5b06c31beb388e0f4c874dbd2dc4f80fcf/tsfc/loopy.py#L436-L451 and we dont since they arent part of C. @dham @mscroggs your thoughts?

As the aim is to be backend (language) agnostic they would be supported if we generated C++: https://en.cppreference.com/w/cpp/numeric/special_functions/cyl_bessel_i

@jhale not that I disagree that they are niche, but the issue was reported here: https://fenicsproject.discourse.group/t/defining-a-fem-function-using-a-conditional-ufl-statement/13730/3?u=dokken and the bessel_i and bessel_k functions do exist in legacy dolfin. I think for...

The default format (pdf) should probably also be optional, as one could store the graphs as png or dot format as well.

@molel-gt I've updated the code to take this into account. Thanks for the feedback.

> There are exceptions, e.g., when we do crystal plasticity examples, computing the Jacobian matrix in Newton iteration could take a significant portion of time (e.g., >80% of the total...

We are aiming to move to numpy 2.0 in DOLFINx (https://github.com/FEniCS/dolfinx/) and we have pyamg in our demos, so would be great to get #407 merged.