I think the lack of speed here is more that it has to serialize the data from disk into a readable format. I assume using the `--grep=` option is faster than piping it through grep because of this