It's working the way it's supposed to. By anchoring the search you are cutting down the number of matches being performed.

Unachored. A 1k file will yield 1k of matches being attempted. Once for every char. position in the file.
Anchored. Will only attempt as many matches as there are lines in the file.

Both are useful in the right context.


edit: PS. additional thought makes me think my math was too simple. I think there would be a lot more than 1k matches being made.