When is code duplication better than a function call?

Question

C++ / C#

Now we are not talking about the speed of development and not about the risks of making a mistake. By efficiency I mean the optimal ratio of execution speed and size of the executable file. For example, instead of calling get / set methods in simple cases, it is more efficient to access the property - as far as I know, competent optimizers do such a substitution. Can you generalize from a theoretical point of view the rule of the situation in which it is more efficient to substitute the body of a function instead of calling it?
Correction to the condition: a function can be called in more than one place, recursion is not required.

Answer 1

If at the moment the function has not been defined in memory (was not required) and the use of the logic of the code of this function is planned strictly once , then it is more reasonable to use the function body itself right away.