Open berquist opened 4 years ago
It is lost to the ages what the actual problem with the iterative solver was (something with separate response vectors rather than a single combined U?), but I had left this draft in the hopes of referencing another paper or two, but clearly that didn't happen. The code has been done the entire time.
A bug where the list holding response vectors (
self.x
) wasn't cleared was hiding the fact thatsolve_dynamic_iterative
was not correct.