We could define two functions which separately compute the contributions to the stress energy tensor associated with the fluid part (which could be shared with the Relativistic Euler), and the magnetic field part. That would likely allow for better code re-use.
We should probably profile the two implementations when deciding. I'd expect there to be little to no performance difference, but it would be good to make sure.
We could define two functions which separately compute the contributions to the stress energy tensor associated with the fluid part (which could be shared with the Relativistic Euler), and the magnetic field part. That would likely allow for better code re-use.