Closed Philosoph228 closed 5 years ago
I wouldn't call it perfection. The matrix inversion is still using the determinant method which is inefficient and due to the number of fops involved also prone to numerical roundoff errors. Eventually this has to be rewritten into something like Gauß-Jordan, or LU or QR decomposition. Heck, even a relaxed Jacobi solver would be better.
Let's finish it to perfection.