I have registered https://t.co/hCAkLbt36C repository on GitHub. On my environment (MacBook Pro 13" 2020 Apple M1), direct tail call threading was faster than switch dispatch. #rustlang
Note that the inline asm for direct threading is only for Aarch64 and only works at opt-level 0. If anyone knows why it does not work at opt-level 1 or higher, please let me know. #rustlang