Hacker News new | past | comments | ask | show | jobs | submit login

Their assembly implementation can be improved, at least when targeting modern hardware.

Here’s an article explaining how to use new instructions to implement large integers multiplication: http://www.intel.com/content/dam/www/public/us/en/documents/...




If you use Rust's equivalent of `-march=native` you get faster code that can target modern hardware. We would have to get another factor-of-two speedup for the maintainance burden of using inline assembly to be worthwhile.

https://gist.github.com/Vurich/5cb83c773e90fc7a463ccb58e1dad...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: