PDEArena

License: MIT · Updated 2026-02-05

PDEArena bundles several spatiotemporal PDE systems (Navier-Stokes, shallow water and more) together with fast differentiable solvers, so you can generate data on the fly and benchmark neural surrogates under identical conditions.

Use cases: Evaluating long-horizon rollout stability of neural operators; data generation without storing huge datasets; comparing surrogate architectures on a shared protocol.

Tags: PDEBenchmarkDifferentiable SolverMicrosoftPyTorch
Related entries
  • PDEBench · Large-scale PDE benchmark covering heat, advection-diffusion and fluid dynamics
  • The Well · Unified benchmark spanning 16 physical domains, dubbed the 'ImageNet moment of p
  • TIDE · 2026 3D turbulent DNS benchmark with 5 tasks and controlled generalization split
  • CFDBench · Multiphysics benchmark for computational fluid dynamics covering internal/extern
  • TU Delft Turbulent Wake Dataset · Classic experimental & simulation dataset of turbulent wakes/cavity flows from T