When I use chinese string to test bechmark, the performance is lower than python method

ashvardanian / StringZilla

Up to 10x faster strings for C, C++, Python, Rust, and Swift, leveraging NEON, AVX2, AVX-512, and SWAR to accelerate search, sort, edit distances, alignment scores, etc 🦖

https://ashvardanian.com/posts/stringzilla/

Apache License 2.0

2.05k stars 66 forks source link

When I use chinese string to test bechmark, the performance is lower than python method #56

Closed ybingHuang closed 7 months ago

ybingHuang commented 10 months ago

python scripts/bench.py --haystack_pattern "你好，StringZilla" --haystack_length 1e9 --needle "你好abce"

stringlizza

ashvardanian commented 10 months ago

@ybingHuang that is possible with your haystack. It fully benefits the Python string and completely disregards my optimizations. I have already tested an improved approach in another repo and will try to release it this month.

ybingHuang commented 10 months ago

I am looking forward to the update ,thank you!

ashvardanian commented 8 months ago

Hi, @ybingHuang! It's not the case anymore on the main-dev. Looking forward to release it soon.