How long will a 2D interFoam (VOF) sloshing study take on this machine? Enter a proposed run and read the anticipated wall-clock to wait. The per-case rate is the machine's own measured µs per cell·timestep (from the CFD study), so the estimate reproduces the real runs — see the validation panel. Use it to scope and review new CFD work before committing the compute.
Model: per-case wall = rate × cells × timesteps (rate = measured, at ~14–16-wide fan-out); batch = ⌈cases / fan-out⌉ waves. Estimates carry ~±30% (adaptive-timestep & free-surface variance). 3D (Nz>1) applies an unmeasured ~1.2× cost/cell penalty — benchmark before relying on it. Running fewer concurrent finishes each case faster (a quieter box is ~35% quicker per case) but fewer in parallel.
| Measured case class | Cells | Timesteps | Fan-out | Measured / case | Predicted / case |
|---|
The per-case rate is calibrated from these classes, so predicted matches measured by construction; the value is extrapolating to new mesh sizes, sim lengths, case counts and concurrency for planning.
Built by scripts/capabilities/build_cfd_runtime_estimator.py from
docs/api/cfd/sloshing-compute.json. — CFD study & provenance →