this post was submitted on 20 Jul 2025
10 points (100.0% liked)

Hacker News

2088 readers
356 users here now

Posts from the RSS Feed of HackerNews.

The feed sometimes contains ads and posts that have been removed by the mod team at HN.

founded 10 months ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] nous@programming.dev 3 points 3 days ago (2 children)

This is such a missleading claim. You cannot get a 100x speed up with handwritten assembly unless the code before what utter shit to start with. Assembly is not the cause for the 100x speedup here, the fact that it was written to use SIMD instructions are the bigger reason for the speedup. You don't require asm to make use of SIMD instructions.

[–] MadMadBunny@lemmy.ca 2 points 3 days ago (1 children)

Maybe using the SIMD instructions, coupled with using Assembly, got the speedup to 100x?

[–] nous@programming.dev 2 points 3 days ago

Even if true, SIMD is doing the heavy lifting here. Probably followed by the fact that almost any rewrite of a code base will result in performance improvements due to the nature of being more familiar with the domain and where the bottle necks are. I would be surprised if the assembly was responsible for more then about 1% of the gains here. So why highlight the fact assembly was used here? It is just missleading. If you want to show how ASM is so much better you need a much better example then this. For all we know the use of ASM could have made things slower and harder to develop. There is just no details at all as to why ASM is beneficial here except some author seems to love it.