Skip to content

Conversation

@pandalee99
Copy link
Contributor

@pandalee99 pandalee99 commented Jan 28, 2025

What does this PR do?

Introduced simdutf to Fury, evaluated the implementation, and then chose a better implementation.

First look at the execution rendering:

1111

Here _SIMD represents the Fury implementation, and _SIMDUTF represents the simdutf implementation.

As you can see, some apis are better, but not all are better, and this may be related to business needs.So I chose some apis that performed better, such as isLatin1

Then following the improved presentation:

2222

Here _FURY represents its own implementation, and _SIMDUTF represents an external library

Related issues

#2013

Does this PR introduce any user-facing change?

  • Does this PR introduce any public API change?
  • Does this PR introduce any binary protocol compatibility change?

Benchmark

@pandalee99
Copy link
Contributor Author

I think FURY's implementations are all better than simdutf's so far

image

@chaokunyang chaokunyang merged commit 261c305 into apache:main May 19, 2025
50 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants