@btime introduces a new scope?

[Gus_Hart]

OK, I found an answer to this elsewhere on discourse, so I’ll answer my own question and hope that it helps someone else:

@btime eats my variables

Welcome! The reason for your issue is that @btime is designed to operate at global scope, so any variables it references need to be globals. That doesn’t work in your case because x is a local, not a global variable. Fortunately, this is easy to work around: you can use $ to interpolate the value of x into the expression being benchmarked: julia> function f(x) @btime $x^2 end f (generic function with 1 method) julia> f(1) 0.024 ns (0 allocations: 0 bytes) Note also that thi…

Variables in the macro need to be global, not local, scope variables. $ can be used to interpolate to the value of the local variable in the @btime call.

Here’s the simple correction to my example:

for k ∈ [5,10,20]
    println(k)
    @btime v1(x[1],$k);
    @btime v2(x[1],$k);
end

and the output:

5
  107.532 ns (2 allocations: 144 bytes)
  63.393 ns (2 allocations: 144 bytes)
10
  187.909 ns (2 allocations: 176 bytes)
  80.880 ns (2 allocations: 176 bytes)
20
  358.810 ns (2 allocations: 256 bytes)
  115.605 ns (2 allocations: 256 bytes)

My “fast” version is sublinear in problem size.