Improving FEniCSx benchmark

[jorgensd]

In

jax-am/jax_am/fem/tests/fenicsx_gold.py

Lines 460 to 461 in e43203f

	# Remark(Tianju): "problem" should be better defined outside of the for loop,
	# but I wasn't able to find a way to assign Dirichlet values on the top boundary.

you state that you have problems with modifying the value of the Dirichlet BC, and hence have to redefine the solver at every time step.

Consider:
https://jsdokken.com/dolfinx-tutorial/chapter2/hyperelasticity.html

which uses a dolfinx.fem.Constant to assign a value to the traction (similarly you can do this for a dirichlet BC), which means that the solver can be defined outside of the loop.

If you need help implementing this I can help you