Strike 2 for the day on the “knocking out stupidity” count.
Tracked down the bugs in my numerical analysis program for the third homework I think. It _really_ helps to not miscalculate the norm of the error in getting O(N^2) behavior out of a preconditioned Richardson method. Similarly, using the right formula to calculate the beta at each iteration in a Chebyshev is useful in getting an O(N) behavior. Urgh, now I have to recreate all of the graphs though.
Just wish I knew why this didn't jump out at me when I was double-checking the code looking for stupidities back when this assignment was due. Good thing the professor more cares that we do it and get it right than get it right on time 😉