I think you should be able to get good results for precalculating the cosines for a set of angles and then using the formula for the cosine of the sum of 2 angl

I think you should be able to get good results for precalculating the cosines for a set of angles and then using the formula for the cosine of the sum of 2 angles combined with the taylor series. The series converges very quickly for very small angles, right?