As a followup, I went and implemented this for fun: https://github.com/bdonlan/branchless-utf8/blob/master/test/decode.s https://github.com/bdonlan/branchless-u

As a followup, I went and implemented this for fun: https://github.com/bdonlan/branchless-utf8/blob/master/test/decode.s https://github.com/bdonlan/branchless-utf8/blob/master/test/... Performance seems to be quite good compared to the implementations from the article (577.333 MB/s vs the author's implementation's 410 MB/s on my hardware)