NSDETimeParallel.jl

Parareal with explicit stopping targets, chunk schedules, and work diagnostics.

First example

using NSDERungeKutta, NSDETimeParallel
problem = Logistic(0.3, (0.0, 2.0))
par = Parareal(RK4(h=0.005), RK4(h=0.05); N=4)
result = solve(problem, par; backend=ThreadsBackend())
@assert result.diagnostics.status in (:converged, :exact)
result.diagnostics.fine_batches

Start with the guided tour, then consult:

• Criteria
• Conventions and cost models
• Backends
• Extensions
• Quick reference
• API

Environment

These pages describe the current development checkout (Julia 1.10 or later). Use compatible versions of the NSDE packages in one project. For local checkouts, develop each package with Pkg.develop(path="/path/to/package"); Pkg.status() reports what will actually load. A registered release may expose an older API. Plotting examples additionally need Plots.

Reference

G. A. Antonucci (2025), Solving chaotic ODEs with time-parallel algorithms, DPhil thesis, University of Oxford. DOI: 10.5287/ora-xoevp0dez. The thesis provides the mathematical background; the API and implementation conventions are specified by these package docs.