> In other words, for common searches (which are short strings), it is much better to spend more time in a vectorized routine than to try to skip bytes. Comple

> In other words, for common searches (which are short strings), it is much better to spend more time in a vectorized routine than to try to skip bytes. Complexity analysis of substring search focuses on the number of comparisons – at least those I saw –, much like sorting, and of course that's not an accurate model at all.