I've been reading "Algorithms, 4th Ed" by Sedgewick & Wayne, and along the way I've been implementing the algorithms discussed in JavaScript.

I recently took the mergesort examples provided in the book to compare top-down and bottom-up approaches... but I'm finding that bottom-up is running faster (I think). See my analysis on my blog. - http://www.akawebdesign.com/2012/04/13/javascript-mergesort-top-down-vs-bottom-up/

I have not been able to find any discussion that says one method of mergesort should be faster than the other. Is my implementation (or analysis) flawed?

Note: my analysis measures the iterative loops of the algorithm, not strictly the array compares/moves. Perhaps this is flawed or irrelevant?

EDIT: My analysis didn't actually time the speed, so my statement about it running "faster" is a bit misleading. I am tracking the "iterations" through the recursive method (top-down) and the for loops (bottom-up) - and bottom-up appears to use fewer iterations.