Description
Polars is one of the fastest-growing DataFrame libraries in the Python ecosystem, and now you can run the same Polars code on the GPU with cudf-polars.
In this "Code With Me" session, Brian Tepera (PM for NVIDIA), William Hill (TME for NVIDIA), and Robin van den Brink (PM for Polars) will walk through why this matters and how to use the GPU engine for Polars.
Demo done on a real financial market dataset.
What you'll learn:
- How to switch an existing Polars LazyFrame query to the GPU engine (.collect(engine="gpu")).
- Which patterns benefit most from GPU acceleration: high-cardinality group_by, sort, join, and windowing.
- How to think about data size and cardinality when deciding between CPU and GPU Polars.
- A live look at a real FSI-flavored workflow on USA stock market data.
Join us live and ask questions throughout the session.