Currently, the access to polynomial data via the variable system is a little cumbersome. We need to know the number of degrees of freedom to access a variable accordingly. This number needs to be passed in, and as of now, we may have different polynomial degrees on different levels. In the future we would like to allow for different polynomial degrees from element to element no matter which level, so this is likely to get more involved. Basically, we always need to look up the number of degrees of freedom for each element.
Luckily, we do have a mechanism to provide information for each element: the variable system. If we had a variable, providing the polynomial degree / number of degrees of freedom for each element, we could look up the appropriate number of degrees of freedom within the get_element routines themselves.
I ran into this with the subsampling for tracking. The subsampling was implemented for data from restart files, where all elements have the same polynomial degree, while this still has to be passed in, the interface would get pretty clumsy if it would need to encompass a list of polynomial degrees for all elements.
For now, I'll stick to the limited functionality with a single degree for all elements. This ticket is a reminder, that this ought to be changed.
Currently, the access to polynomial data via the variable system is a little cumbersome. We need to know the number of degrees of freedom to access a variable accordingly. This number needs to be passed in, and as of now, we may have different polynomial degrees on different levels. In the future we would like to allow for different polynomial degrees from element to element no matter which level, so this is likely to get more involved. Basically, we always need to look up the number of degrees of freedom for each element.
Luckily, we do have a mechanism to provide information for each element: the variable system. If we had a variable, providing the polynomial degree / number of degrees of freedom for each element, we could look up the appropriate number of degrees of freedom within the get_element routines themselves.
I ran into this with the subsampling for tracking. The subsampling was implemented for data from restart files, where all elements have the same polynomial degree, while this still has to be passed in, the interface would get pretty clumsy if it would need to encompass a list of polynomial degrees for all elements.
For now, I'll stick to the limited functionality with a single degree for all elements. This ticket is a reminder, that this ought to be changed.