Skip to content

Fix and update AdaptiveSDE benchmarks (#126) #1605

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
jitendravjh wants to merge 8 commits into
base: master
Choose a base branch
from
+334 −269
Open

Fix and update AdaptiveSDE benchmarks (#126) #1605

Show file tree
Hide file tree
Changes from 6 commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
0e2a8d6
Fix and update AdaptiveSDE benchmarks (#126)
jitendravjh Jun 28, 2026
c69afef
Route qmax through the step-size controller in qmaxDetermination
jitendravjh Jul 2, 2026
be28a34
Use SOSRI for the Oval2 qmax test
jitendravjh Jul 2, 2026
669a849
Address review: transpose RSwM legend labels and align Oval2 qmax tol…
jitendravjh Jul 5, 2026
b597138
Tune Oval2 qmax controller to gamma=0.7 for zero adaptive failures
jitendravjh Jul 6, 2026
c0bf9f8
Force dtmin and cap maxiters on oval2 qmax sweep
jitendravjh Jul 21, 2026
50c97ea
Switch oval2 qmax test to SROCKEM
jitendravjh Jul 28, 2026
048d11a
Restore SRIW1 on oval2 qmax test and bump DiffEqNoiseProcess to 5.34.0
jitendravjh Aug 5, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
43 changes: 15 additions & 28 deletions benchmarks/AdaptiveSDE/AdaptiveEfficiencyTests.jmd
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,25 +4,14 @@ author: Chris Rackauckas
---

```julia

using Distributed
addprocs(2)
using StochasticDiffEq, SDEProblemLibrary, DiffEqNoiseProcess, DiffEqBase, Plots
import SDEProblemLibrary: prob_sde_additive,
prob_sde_linear, prob_sde_wave

p1 = Vector{Any}(undef, 3)
p2 = Vector{Any}(undef, 3)
p3 = Vector{Any}(undef, 3)

@everywhere begin
using StochasticDiffEq, SDEProblemLibrary, DiffEqNoiseProcess, Plots,
ParallelDataTransfer
import SDEProblemLibrary: prob_sde_additive,
prob_sde_linear, prob_sde_wave
end

using StochasticDiffEq, SDEProblemLibrary, DiffEqNoiseProcess, Plots, ParallelDataTransfer
import SDEProblemLibrary: prob_sde_additive,
prob_sde_linear, prob_sde_wave

probs = Matrix{SDEProblem}(undef, 3, 3)
## Problem 1
prob = prob_sde_linear
Expand Down Expand Up @@ -59,7 +48,8 @@ Ns = [17, 23,
17]
```

Timings are only valid if no workers die. Workers die if you run out of memory.
The Monte Carlo runs use multithreaded ensembles (`EnsembleThreads()`), so start Julia
with multiple threads (e.g. `julia -t auto`) for parallel timing.

```julia
for k in 1:size(probs, 1)
Expand All @@ -76,16 +66,15 @@ for k in 1:size(probs, 1)

#Compile
prob = probs[k, 1]
ParallelDataTransfer.sendto(workers(), prob = prob)
monte_prob = EnsembleProblem(prob)
solve(monte_prob, SRIW1(), dt = 1/2^(4), adaptive = true,
solve(monte_prob, SRIW1(), EnsembleThreads(), dt = 1/2^(4), adaptive = true,
trajectories = 1000, abstol = 2.0^(-1), reltol = 0)

println("RSwM1")
for i in (1 + offset):(N + offset)
tols[i - offset, 1] = 2.0^(-i-1)
msims[i - offset] = DiffEqBase.calculate_monte_errors(solve(monte_prob, SRIW1(),
trajectories = 1000, abstol = 2.0^(-i-1),
msims[i - offset] = DiffEqBase.calculate_ensemble_errors(solve(monte_prob, SRIW1(),
EnsembleThreads(), trajectories = 1000, abstol = 2.0^(-i-1),
reltol = 0, force_dtmin = true))
elapsed[i - offset, 1] = msims[i - offset].elapsedTime
medians[i - offset, 1] = msims[i - offset].error_medians[:final]
Expand All @@ -95,15 +84,14 @@ for k in 1:size(probs, 1)
println("RSwM2")
prob = probs[k, 2]

ParallelDataTransfer.sendto(workers(), prob = prob)
monte_prob = EnsembleProblem(prob)
solve(monte_prob, SRIW1(), dt = 1/2^(4), adaptive = true,
solve(monte_prob, SRIW1(), EnsembleThreads(), dt = 1/2^(4), adaptive = true,
trajectories = 1000, abstol = 2.0^(-1), reltol = 0)

for i in (1 + offset):(N + offset)
tols[i - offset, 2] = 2.0^(-i-1)
msims[i - offset] = DiffEqBase.calculate_monte_errors(solve(monte_prob, SRIW1(),
trajectories = 1000, abstol = 2.0^(-i-1),
msims[i - offset] = DiffEqBase.calculate_ensemble_errors(solve(monte_prob, SRIW1(),
EnsembleThreads(), trajectories = 1000, abstol = 2.0^(-i-1),
reltol = 0, force_dtmin = true))
elapsed[i - offset, 2] = msims[i - offset].elapsedTime
medians[i - offset, 2] = msims[i - offset].error_medians[:final]
Expand All @@ -112,15 +100,14 @@ for k in 1:size(probs, 1)

println("RSwM3")
prob = probs[k, 3]
ParallelDataTransfer.sendto(workers(), prob = prob)
monte_prob = EnsembleProblem(prob)
solve(monte_prob, SRIW1(), dt = 1/2^(4), adaptive = true,
solve(monte_prob, SRIW1(), EnsembleThreads(), dt = 1/2^(4), adaptive = true,
trajectories = 1000, abstol = 2.0^(-1), reltol = 0)

for i in (1 + offset):(N + offset)
tols[i - offset, 3] = 2.0^(-i-1)
msims[i - offset] = DiffEqBase.calculate_monte_errors(solve(monte_prob, SRIW1(),
adaptive = true, trajectories = 1000, abstol = 2.0^(-i-1),
msims[i - offset] = DiffEqBase.calculate_ensemble_errors(solve(monte_prob, SRIW1(),
EnsembleThreads(), adaptive = true, trajectories = 1000, abstol = 2.0^(-i-1),
reltol = 0, force_dtmin = true))
elapsed[i - offset, 3] = msims[i - offset].elapsedTime
medians[i - offset, 3] = msims[i - offset].error_medians[:final]
Expand All @@ -137,7 +124,7 @@ end
```julia
gr(fmt = :svg)
lw=3
leg=String["RSwM1", "RSwM2", "RSwM3"]
leg=["RSwM1" "RSwM2" "RSwM3"]

titleFontSize = 16
guideFontSize = 14
Expand Down
Loading
Loading