Ask HN: numerical source code worth studying

Netlib (http://www.netlib.org/) is the home of LAPACK and BLAS and many, many other classic, robust numerical code, almost all in Fortran 77, and would typically need advanced background in numerical analysis to understand the details.

In any case, if you want to do numerical analysis, it is better to use existing code than trying to write your own.

There is a lot of open-source numerical processing code, most but not all Python, in Sage:

http://www.sagemath.org/

Also, there is quite a lot of symbolic algebra processing code in the project. This might not be your primary interest.

Numpy, Scipy, Octave, Quantlib