mkleen opened a new pull request, #26954: URL: https://github.com/apache/datafusion/pull/26954
## Which issue does this PR close? - Closes None ## Rationale for this change `normalize_floats` was added in #26595. It runs on every float key column before the column is compared, hashed or grouped, so it is on the hot path for joins, aggregations and sorts with float keys. This benchmark is the baseline for a follow-up PR that speeds up the scan. ## What changes are included in this PR? - A new `float_normalization` criterion bench in `datafusion-common` - It measures `normalize_floats` on one batch of 8192 random values (fixed seed), for both `Float64` and `Float32` with the following combinations: - `clean`: nothing to normalize, so the input is returned unchanged. - `neg_zero_last`: a single `-0.0` in the last row, so the whole column is scanned and then copied. Run it with: ```bash cargo bench -p datafusion-common --bench float_normalization ``` ## What is the testing strategy for this PR? It's only a benchmark ## Are there any user-facing changes? No. -- 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]
