You could actually use SIMD instructions to detect sequences of bytes with the top bit cleared, thus allowing bulk copies of ASCII text without a per-byte loop.

Another optimization would be to take advantage of the fact that codepoint usage tends to cluster around the language of the text. So if you detect usage of 3-byte encodings, chances are you'll continue encountering only 3-byte encodings, with the odd ASCII or emoji codepoints. This opens up even more state machine possibilities.

> So if you detect usage of 3-byte encodings, chances are you'll continue encountering only 3-byte encodings, with the odd ASCII or emoji codepoints.

Depends on what kind of text you're processing. Many languages use the ASCII range for spaces/newlines, digits, and punctuation.

FWIW I believe the second optimization is at odds with the first. It’s really hard to do this kind of conditional historical check in SIMD.

This is one of the optimizations from ClickHouse - detect ASCII and go a fast path: https://github.com/ClickHouse/ClickHouse/blob/6cde32de1a2463...