There was a need to write a quick calculation of Sin and Cos. The basis for the calculations took the decomposition of the Taylor series. I use it in 3D systems (OpenGL and a graphical library of my own development). Unfortunately, it’s not possible to reduce the series “perfectly” for Double, but this is offset by good acceleration. The code is written in the assembler built into Delphi XE6. Used by SSE2.
Not suitable for scientific computing, but quite suitable for use in games.
There is enough accuracy to cover the digit capacity of the Single number, which is used