MS-COFD-2DSL dataset production

2D steady laminar flow around obstacles · OpenFOAM on the KUACC cluster

Work done

–

Estimated finish

–

Jobs

–

Finished, last hour

–

Disk quota

–

Compute used

–

Runs per mesh level

Level	Progress	Done	Running	Not converged	Failed	Mean h / run	Last 24 h	Core-h used	Core-h left (est.)

Green: finished · blue: running. The 0.5 and 0.25 mm meshes are run only for cases whose 1 mm solution converged. Not converged: cases of the dataset whose latest run stopped at its iteration cap; dropped and replaced cases, and runs that converged when run again, are listed under Problems. Core-hours: solver time (meshing, arrays and attributes add ~5–10 %). Left: estimated for every remaining case from its blockage ratio, since the solver time grows with it (0.25 mm: ~2.5 h at blockage 0.12, ~12 h at 0.40). The fit comes from the pilot until half of a level is done, then from the finished runs; a plain mean would be too low because the fast runs finish first.

The race to 10⁻⁶

Every running job is a runner, drawn as its obstacle. A runner's position is how far the solver residual has fallen toward the 10⁻⁶ finish line (log scale). Each lane's axis starts at its slowest runner, so it zooms in where the runners are. Between updates, runners keep moving at their measured pace. Hover or tap a runner for details.

Closest to the finish

Fastest pace on each mesh

Longest on the road

Needs attention

Finish line

Milestones

Progress over time

Nodes in use

Problems

Cases without a steady solution

Their 1 mm run stopped at the iteration cap: the flow is not steady. They are replaced from reserve geometries at the end, so every class keeps 300 converged cases.