It looks like converting surfaces to triangles for display can take a lot of time. SolveSpace doesn't make use of multi-core CPUs as far as I can tell. One place we might try this is in the amazingly short function:

void SShell::TriangulateInto(SMesh *sm) {

    SSurface *s;
    for(s = surface.First(); s; s = surface.NextAfter(s)) {
        s->TriangulateInto(this, sm);
    }
}

All shells will have multiple surfaces. I'm not sure that function call can be done in parallel. Maybe we could have a look. If so, we'd just create a vector of triangle lists and use OpenMP parallel for (or equivalent?) to go over handing each surface its own list to populate and then concatenate the triangle lists after.

This hinges on weather that is a pure function. If so, what's the preferred way to do things in parallel? I'm not aware of any place SS does that.