The original code was bundled with CGAL, between versions 3.6 (March, 2010) and 4.3 (October, 2013). For CGAL 4.4, I removed it because I rewrote the algebraic kernel that used it. I release it because I hope the code is useful for someone. If you are interested, don't hesitate in contacting me.
The implementation the well-known Euclidean algorithm to compute GCD. However, big numbers make the size of the coefficients explode. We use then modular arithmetic. We compute images of the input polynomials modulo some primes and compute then images of these modular polynomials. Then, we use Chinese lifting to recover the result.
You can compile using GNU make. Edit makefile and config to suit your needs.
The code is tested automatically via Travis CI using GCC and Clang, both under GNU/Linux and Mac OS. Tests are triggered when a new commit is pushed. The code should also compile on MSVC, although I did not test it.
- von Zur Gathen and Gerhard, Modern Computer Algebra, 2nd edition, Cambridge, 2003.
- Geddes, Czapor and Labahn, Algorithms for Computer Algebra, Springer, 1992.
- Zippel, Effective Polynomial Computations, Springer, 1993.
The code is released under the LGPLv3. See files COPYING.LESSER and COPYING in the bundle for more information. Originally, the code was released under the LGPLv2.