Are there functional or performance differences between

• myfunction(x::Real), and
• myfunction(x::T) where {T<:Real}?

In this case, Real is an abstract type which obviously has concrete subtypes like Float64 and Int.

Are there reasons to prefer one versus the other?

CodePudding user response：

The biggest difference is that you can refer to T in the function definition. The other difference is that for Functions and Varargs (but no other types, myfunction(x::T) where {T} forces specialization.

Other than that, they are exactly the same.