What is the Compute 3D Model? And what is it for?

Compute 3D Model is the core component of Cognitive Design.

It computes and displays geometry coming from any BRep, Mesh, or FRep input. Almost every workflow requires it: it evaluates the input chain, generates the resulting geometry, and provides a consistent viewer to inspect each design step before passing the result to downstream tools.

image-20251212-031450.png

What it is for

Inputs

Computation

When the model is computed, implicit geometry is sampled on a three-dimensional grid, and the step of that grid decides which features survive. Anything much thinner than the step is not represented at all, and sharp edges are progressively rounded off as the step grows.

The rule to work by: the sampling step should be at most half the smallest feature you care about. A 1 mm rib needs a step of 0.5 mm or finer; a 0.4 mm lattice wall needs 0.2 mm.

<aside> ⚠️

The named presets are not a fixed size in millimetres. Each one spreads a fixed sampling budget across the bounding box of your model, so the step it produces scales with the size of the part. Fine on a 50 mm bracket is far finer than Fine on a 2 m beam. This is why the same preset can look perfect on one part and lose detail on another — when a specific feature size has to be resolved, use Custom and set the step explicitly.

</aside>

Cost. Halving the sampling step multiplies the amount of data by roughly eight, so compute time and memory rise steeply. Refine only when you need to, and only at the end.

Visualization

After computation, the 3D model appears in the viewer. You can access: