allocating linearalgebra

[rveltz]

Hi,

I am surprised by the following behavior. Can you please tell me if this is expected?

Thank you,

Best regards,

using BenchmarkTools

xce0 = ArrayPartition(zeros(2),[0.])
xcde0 = copy(xce0)

function foo(y, x)
	y .= y .+ 2 .* x
end

@btime foo($xcde0, $xce0)

which gives

julia> @btime foo($xcde0, $xce0)
  30.138 ns (3 allocations: 48 bytes)

[ChrisRackauckas]

It's 48 bytes, so it could just be the fact that you're returning it. What happens if you add nothing to the end?

[rveltz]

Nothing :(

What's more:

function foo(y, x)
	y .= y .+ 2 .* x .+ 3 .* x
	nothing
end

@btime foo($xcde0, $xce0)

gives 38.338 ns (4 allocations: 64 bytes)

and

julia> function foo(y, x)
               y .= y .+ 2 .* x .+ 3 .* x .+ x
               nothing
       end
foo (generic function with 1 method)

julia> @btime foo($xcde0, $xce0)
  46.252 ns (5 allocations: 80 bytes)

I can see a pattern...

[ChrisRackauckas]

Uhh, can't say I have an explanation for that. I wonder if there's something in the broadcast definition we should be avoiding. Whatever it is, this will take time.