jhorstmann commented on PR #9047:
URL: https://github.com/apache/arrow-rs/pull/9047#issuecomment-3728127841
I also benchmarked the `zmij` and `teju` crates in our code base, before
seeing this PR, and `zmij` seems to be 2x faster than the current arrow code
(`TO_STRING/FLOAT` is directly using the arrow `cast` kernel):
```
TO_STRING/FLOAT time: [2.2333 ms 2.2349 ms 2.2365 ms]
thrpt: [29.302 Melem/s 29.324 Melem/s 29.345
Melem/s]
TO_STRING/FLOAT_TEJU time: [1.7545 ms 1.7663 ms 1.7816 ms]
thrpt: [36.784 Melem/s 37.104 Melem/s 37.353
Melem/s]
TO_STRING/FLOAT_ZMIJ time: [979.91 µs 985.56 µs 993.57 µs]
thrpt: [65.960 Melem/s 66.496 Melem/s 66.880
Melem/s]
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
