NVIDIA shows how to run Polars on the GPU with cudf-polars
NVIDIA Developer · youtube · 2026-07-21
NVIDIA is showing how to run existing Polars queries on the GPU with cudf-polars.
- The session demonstrates switching a LazyFrame query to GPU execution with .collect(engine="gpu").
- It focuses on patterns that benefit most from acceleration: high-cardinality groupby, sort, join, and window functions.
- The demo uses a real financial market dataset and explains how to decide between CPU and GPU based on data size and cardinality.
More from Infra
- Tabul AI launches Metal TreeSHAP to speed up Shapley values on Apple silicon — Scobleizer · 2026-07-22
- DeepSeek-V4-Flash tops out at 770 tok/s on one B300 in a vLLM batch test — Moreh · 2026-07-22
- NVIDIA starts shipping 102.4 Tbps Spectrum-6 switches for Vera Rubin AI factories — nvidia · 2026-07-22
- Apple publishes SOC 3 audit reports for Private Cloud Compute — throwfaraway4 · 2026-07-22
- Reddit GPU renters say existing platforms only give you two of three: code, recovery, fair billing — legendpizzasenpai · 2026-07-22
- The Sandboxing Manifesto: Secure Execution Environments for Agents — spirosoik · 2026-07-22