multithreading

Problem
I am returning to WebGL work after some time away (with Vulkan) so may be misinterpreting something here, but...
The WebGL example loops through the render_gl function every frame. Therefore it's recompiling the shaders, linking them, all of this work that doesn't need to be repeated in the render loop.

Steps To Reproduce
Steps to reproduce the behavior:
Start the app as

FEATURE REQUEST:

In CannyLab/tsne-cuda#8, the possibility of using a custom NN matrix is discussed and noted to be 'easy' to implement.
DavidMChan: " It would be easy to add the ability to pass in a sparse nearest neighbors matrix, however it becomes more complicated if you want to extract the nearest neighbors from a pre-computed distance matrix."

It would be a s