This is definitely not yet ready for merging, but I have reached the point where this new approach offers comparable performance to the old (when applied to the expensive compute_jhj_jhr function). I need to propagate these changes to all the solvers and check that I haven't broken anything.
This is definitely not yet ready for merging, but I have reached the point where this new approach offers comparable performance to the old (when applied to the expensive
compute_jhj_jhr
function). I need to propagate these changes to all the solvers and check that I haven't broken anything.