Yes, the latter will be faster.
Directory access is a linear search on an unsorted pile-of-stuff.
On reasonably sized directories that's not a big deal, but if you put thousands of files in a directory, performance will suffer.
If you get really anal-retentive about this issue you'll wind up with a filesystem that amounts to a sorted binary tree. But there are practical limits to path length in various Unixes (Unices?).
So just do something practical, like your example.
