Is it possible to precompile a function in Julia?

General Usage Performance
,
1

Hi there,

I’m trying to make my Julia code as fast as possible and a friend of mine recommended using precompile. I have read about it but I’m not sure how to proceed, do I precompile the package I’m using o the function I’m creating? This is an example of my code:

`using GLM, Polynomials

function polynomial(k)

x_pol = Polynomials.fit(x, y, k)

return x_pol

end

precompile(polynomial, (Int64, ))`

I haven’t seen any changes in performance of the code when I add precompile(polynomial, (Int64, )) so I’m not sure how to proceed.

Thank you very much.

2

  1. Precompilation only speeds up code loading (i.e. startup time), not runtime performance. All Julia functions are compiled the first time you call them (with a particular set of argument types), and thereafter are fast.

  2. Precompilation is something that happens for code that you put into a package, which you should do for any longer-term development effort in Julia. See also Best practise: organising code in Julia - #2 by stevengj

3 Likes
3

Okey, thank you very much.