I for one find things like these extremely interesting. The more there's analysis about performance implications of certain ways of doing the task the better. T

I for one find things like these extremely interesting. The more there's analysis about performance implications of certain ways of doing the task the better. The only thing I'm left missing here is to actually look into the profiling information(branches, caches, instructions per cycle etc.), especially on the compiler generated assembly to see what exact instruction sequences are to blame, and if there are ways to save cycles here or there. If only there were more things like this in the internets with deep analysis about performance of a given piece of code... :)