OpenGL Mitigate Floating Point Errors

I'm sure some of you will have noticed by now, but previously floating point rounding errors would accumulate in the shaders. If you haven't encountered this yet checkout develop, zoom in really far on a shader visualisation (sunburst, basic tree) (at least till you get the warning message) and translate the viewport a lot, just translating irradically in a lot of directions should do the trick. Now manually (manually because the reset transformations function already corrects these floating point errors) zoom out again and observe some interesting results. This PR mitigates this issue so the effects are unnoticeable (because solving it is of course impossible).