NVIDIA / NVIDIA/cudf

[Story] Re-enable regular PDSH benchmarking with cudf.pandas

Open
#23,629 0 comments 0 reactions 1 assignee Claimed by @mroeschke View on GitHub
cudf.pandas
Dominant language
C++
Stars
9.8k
Forks
1.1k
Avg merge
3d 6m
Merged PRs (30d)
278

Description

We've disabled regular runs with `python/cudf/cudf/pandas/_benchmarks/pdsh.py` with SF-50/100 ~4 months ago because we observed hangs(?) in some queries. Now that cudf.pandas exclusively supports pandas 3, these hangs may not exist anymore.

To close out this issue:

* Ensure all 22 queries pass and are validated with default configurations. Ensuring the queries don't fall back to pandas is out of scope
* Ensure all queries support the "default" generated TPCH data, namely parquet files with `date32` and `decimal` columns

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.