kamille
kamille
It is poc version of #12996
The latest benchmark numbers: ``` -------------------- Benchmark clickbench_1.json -------------------- ┏━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓ ┃ Query ┃ main ┃ vectorize-append-value ┃ Change ┃ ┡━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩ │ QQuery 0 │ 0.67ms │ 0.68ms │ no change...
The logic is a bit complex, I plan to finish and do benchmark for it today.
It is happy to see that the vectorized approach is promising! The rest work is to find why it get `q28` slower and fix tests. ``` -------------------- Benchmark clickbench_1.json --------------------...
I profile to get flamegraph of `q28` and found it actually not go the path in this pr. The slower may be due to some enviorment reason in my local,...
I think this pr is ready now!
> Amazing!!! > > > 1.24x slower > > Worth profiling / checking this regression before merging I think 🤔 I plan to try to check it today or tomorrow,...
> Is it possible to somehow unify GroupValuesColumn and VectorizedGroupValuesColumn ? 🤔I think It can unify simply, `VectorizedGroupValuesColumn::scalarized_intern` is similar as `GroupValuesColumn::intern`. But its logic is much more complex, I...
> > 🤔 I personally prefer the second one? What do you think about it @alamb ? > > I think this makes sense -- thank you Have unified the...
@Xuanwo hi, have a moment to review?