While I was writing my book I uncovered a lot of useful information that I used to make upgrades to my radar cross section (RCS) code
Serenity. Among these improvements were:
Greatly improved the accuracy of the numerical integrations contributing to near-matrix elements in EFIE and MFIE
Added far better Conjugate Gradient Squared (CGS), Biconjugate Gradient Stabilized (BiCG-Stab) and GMRES(k) iterative solvers. I then immediately threw away the Conjugate Gradient solver I had been using.
Added a suite of preconditioners which greatly improve the performance of the MLFMA iterative solver, to the point of ZOMG.
Having now tested the code for a month on many different cases, it is safe to say that Serenity is now one of the most powerful radar cross section solvers in the world. If you have a big enough computer, I doubt there is an RCS problem that Serenity would not crush. Utterly.