integrate some ideas from glslEditor

[gre]

https://github.com/patriciogonzalezvivo/glslEditor

• profiling glsl instructions to identify bottleneck is neat. we could just weight each glsl line running IN transition() function (i'm thinking a relative number from 1 to inf might be better than a % of the total? or maybe the value of total% * nbOfLine). i think it works by replacing a placeholder value to each line and doing bunch of rendering to evaluate.