SciML/OrdinaryDiffEq.jl

Avoid allocations for interpolator evaluation

Open

#1,270 opened on Sep 15, 2020

View on GitHub
 (6 comments) (0 reactions) (0 assignees)Julia (262 forks)batch import
help wantedperformance

Repository metrics

Stars
 (656 stars)
PR merge metrics
 (Avg merge 3d 12h) (74 merged PRs in 30d)

Description

Hi,

I have noticed that the interpolation of a solution allocates. This is a MWE:

using OrdinaryDiffEq
using BenchmarkTools

function f1(du, u, p, t)
  du[1] = u[1]
end
prob1 = ODEProblem{true}(f1, ones(1), (0.0, 1.0), nothing)
sol1 = solve(prob1, Tsit5())

f2(u, p, t) = u
prob2 = ODEProblem{false}(f2, 1.0, (0.0, 1.0), nothing)
sol2 = solve(prob2, Tsit5())

out = zeros(1)

Then:

julia> @btime $sol1(0.1)
  1.181 μs (18 allocations: 416 bytes)
1-element Array{Float64,1}:
 1.1051709180988993

julia> @btime $sol1($out, 0.1)
  1.143 μs (17 allocations: 320 bytes)
1-element Array{Float64,1}:
 1.1051709180988993

julia> @btime $sol2(0.1)
  1.267 μs (25 allocations: 1.33 KiB)
1.1051709180988993

julia> @btime $sol2($out, 0.1)
  1.185 μs (24 allocations: 1.31 KiB)
1-element Array{Float64,1}:
 1.1051709180988993

Thanks!

Contributor guide

Research direction
Profile the interpolation function calls to identify allocation sources, then refactor to use in place operations or pre allocated buffers.
Tech stack
julia
Domain
backend
Issue type
Bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
2
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Julia
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
65