perf: PQ performance - Githubissues

lancedb / lance

Modern columnar data format for ML and LLMs implemented in Rust. Convert from parquet in 2 lines of code for 100x faster random access, vector index, and data versioning. Compatible with Pandas, DuckDB, Polars, Pyarrow, and PyTorch with more integrations coming..

https://lancedb.github.io/lance/

Apache License 2.0

3.97k stars 224 forks source link

perf: PQ performance #3149

Open BubbleCal opened 1 day ago

BubbleCal commented 1 day ago

Now we are optimizing the PQ performance, here to track all potential bottlenecks:

[ ] transforming: now we handle 4bit PQ case while transforming, which requires to collect the intermediate results
[ ] cache locality: the distance calculating can be optimized by better access pattern to distance table

wjones127 commented 1 day ago

Have we fixed either of these yet?

BubbleCal commented 1 day ago

Have we fixed either of these yet?

[ ] perf: presize KeepFiniteVectors transform #2838

[ ] perf: PQ transform re-allocates a lot #2837

not yet