Another helping you get help suggestion: you might provide more context what an "RK4 for a vectorial function" is, because I for sure have no clue what you're referring to off hand. If this is some common algorithm for some specific problem domain that's all well and good, but the peanut gallery here is (like myself) probably completely unfamiliar with what you're talking about.

I mean a wild guess and wolfram alpha query turns up Runge-Kutta Method which looks similar if I squint and scroll around. But giving that context out up front rather than making people trying to help do the work first helps us (well, someone here maybe) help you.