Can you please explain me why in the following code the second broadcasting results in 3 allocations? using BenchmarkTools u = rand(100) k1 = similar(u) k2 = similar(u) k3 = similar(u) k4 = similar(u) @btime @. $u = $u + 1 * ($k1 + $k2 + $k3) @btime @. $u = $u + 1 * ($k1 + $k2 + $k3 + $k4) @btime…

Up. Meanwhile, in Julia 1.4 there are already 4 allocations in the above code. Just to highlight the problem, the following code using BenchmarkTools # import DifferentialEquations function foo(u, k1, k2, k3, k4) for i=1:1000 @. u = u + 1.0 * (k1 + k2 + k3 + k4) end return no…

It seems there is some heuristic that is getting hit because removing the +k4 makes it stop allocating.

Do you have any ideas where I can start to dig in?

Doing julia> function foo(u, k1, k2, k3, k4) @. u = u + 1.0 * (k1 + k2 + k3 + k4) return nothing end foo (generic function with 2 methods) julia> @code_warntype optimize=true foo(u, k1, k2, k3, k4) I see a 8 ─── %24 = invoke Base.Broadcast.combine_axes(1.0::Float64,…

[image] fedoroff: The above code is typical for various Runge-Kutta methods. So far in my realizations of Runge-Kutta methods of order higher than 3 I inevitably see these broadcasting allocations. Can you, please, help me to understand what is going on. Especially, why the import of Differentia…

Do you have a PR that upstreams those?

There wasn’t before, but https://github.com/JuliaLang/julia/pull/35260 is it. Someone should probably double check that those changes are sufficient to get rid of these allocations.

[image] fedoroff: import Base.Broadcast: Broadcasted @inline combine_axes(A, B) = broadcast_shape(broadcast_axes(A), broadcast_axes(B)) @inline check_broadcast_axes(shp, A::Union{Number, Array, Broadcasted}) = check_broadcast_shape(shp, axes(A)) You’re shadowing and not overloading: @inline …

Thank you. Now it works: using BenchmarkTools @inline Base.Broadcast.combine_axes(A, B) = Base.Broadcast.broadcast_shape(Base.Broadcast.broadcast_axes(A), Base.Broadcast.broadcast_axes(B)) function foo(u, k1, k2, k3, k4) for i=1:1000 @. u = u + 1.0 * (k1 + k2 + k3 + k4) end re…

Strange allocations during broadcasting

General Usage Performance

ChrisRackauckas March 25, 2020, 6:02pm 10

You’re shadowing and not overloading:

@inline Base.Broadcast.combine_axes(A, B) = broadcast_shape(broadcast_axes(A), broadcast_axes(B))
@inline Base.Broadcast.check_broadcast_axes(shp, A::Union{Number, Array, Broadcasted}) = check_broadcast_shape(shp, axes(A))

Topic		Replies	Views
Custom structs in JuliaDiffEq: In-place broadcasting results in allocation and slowdown General Usage performance	7	794	December 14, 2021
Unknown allocation Performance	5	662	February 10, 2020
Why does broadcast operation allocate? Performance	3	1368	July 28, 2021
Why is broadcasting allocating so much memory? Performance	8	871	May 21, 2020
Unexpected allocations in looped vs broadcasted functions on tuples of arrays Performance	4	575	March 12, 2020

Strange allocations during broadcasting

Related topics