I have written a bunch of mathematical software in C#, mainly to do some experiments in Number theory (e.g. arbitrary precision integer arithmetic).

I would like to make optimal use of my Intel I7 64-bit machine and be able to compute things like

(a * b) % c (a, b, c being 64-bit integers) using native machine instructions without having to split the numbers into multiple pieces.

Questions:

- what is the best way to extend C# with assembler code for very high performance?

- what dies the MKL provide in this respect