optimization release

This release includes the following changes:

 1) some very significant performance optimizations/on some larger glb files, load times went from 8s to 1.5s after we disabled exception handling
 2) also memory usage for large 3d models was improved significantly/found a bug where vertex arrays were unnecessarily duplicated before they were sent to gpu/duplicate faces were not visible at all, but just did eat perf unnecessarily and hog memory like hell.