Some approaches to speed up collision detection

Processes

Launcing standalone processes, each loads the whole model but checks a subset of expressIDs => full CPU utilization

Simple

Standard collision detection code

web-ifc

Trying to load the geometry directly out of web-ifc => No speed up

worker

Spawning worker threads and serializing the mesh contents between them. Doesnt engage CPU fully

Other tests

• gltfExport: Checking how a conversion from web-ifc to gltf could look like: Not working right now
• objExport: Working implementation of generating .obj files from web-ifc geometry.