Dandandan opened a new issue, #21843: URL: https://github.com/apache/datafusion/issues/21843
### Is your feature request related to a problem or challenge? Currently we use `hash % num_partition` to divide rows among partitions, based on the hash. However the % operation is slow on all platforms (especially x86, but ARM as well) and makes `RepartitionExec` slower. ### Describe the solution you'd like We can consider using fastrange https://github.com/lemire/fastrange instead. ### Describe alternatives you've considered We can apply strength reduction (https://en.wikipedia.org/wiki/Strength_reduction) or the crate https://docs.rs/strength_reduce/latest/strength_reduce/ to make the `hash % partition` calculation faster, as partition is a constant over the batch session config. ### Additional context _No response_ -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
