This time is was due to different accuracy of floating point computation, now it uses does a comparison a bit different to avoid this. Also changed the vectoquat function to be threadsafe.
36 KiB
36 KiB