merge release/26.06 into main

[jameslamb]

Description

Closes #22709

Fixes forward-merger conflicts.

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[jameslamb]

The conflict resolutions were straightforward and non-controversial, cancelling CI to save resources and admin no-squash merging this.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: c2208d03-0cdd-4e44-9329-d9dcb4376067

📥 Commits

Reviewing files that changed from the base of the PR and between 8f0f5d0 and abffe7f.

⛔ Files ignored due to path filters (4)

• docs/cudf/source/_static/polars_pdsds_sf1k.png is excluded by !**/*.png
• docs/cudf/source/_static/polars_pdsds_sf3k.png is excluded by !**/*.png
• docs/cudf/source/_static/polars_pdsh_sf1k.png is excluded by !**/*.png
• docs/cudf/source/_static/polars_pdsh_sf3k.png is excluded by !**/*.png

📒 Files selected for processing (10)

• conda/environments/all_cuda-129_arch-aarch64.yaml
• conda/environments/all_cuda-129_arch-x86_64.yaml
• conda/environments/all_cuda-132_arch-aarch64.yaml
• conda/environments/all_cuda-132_arch-x86_64.yaml
• conda/recipes/cudf/recipe.yaml
• dependencies.yaml
• docs/cudf/source/cudf_polars/index.md
• python/cudf/pyproject.toml
• python/dask_cudf/pyproject.toml
• python/pylibcudf/pyproject.toml

---

📝 Walkthrough

Summary by CodeRabbit

• Chores

  • Updated CuPy dependency version constraints across conda environments, conda recipes, and Python project configurations to exclude additional incompatible package versions.

• Documentation

  • Refreshed Polars GPU engine documentation with new benchmark narratives, comparative performance information, and single-GPU to multi-GPU scaling details.

Walkthrough

This forward-merge from the release branch adds CuPy version 14.1.0 exclusions to conda environments, conda recipes, and Python project dependencies across the cudf ecosystem, plus updates Polars GPU engine benchmark documentation with new table-based figures and simplified narrative.

Changes

CuPy 14.1.0 Version Exclusion and Documentation Updates

Layer / File(s)	Summary
Conda environment and recipe CuPy 14.1.0 exclusions conda/environments/all_cuda-129_arch-aarch64.yaml, conda/environments/all_cuda-129_arch-x86_64.yaml, conda/environments/all_cuda-132_arch-aarch64.yaml, conda/environments/all_cuda-132_arch-x86_64.yaml, conda/recipes/cudf/recipe.yaml	Four conda environment specs (CUDA 12.9 and 13.2 for both aarch64 and x86_64) and the cudf conda recipe each add !=14.1.0 to existing cupy>=13.6.0,!=14.0.0 constraints.
Python project dependency CuPy 14.1.0 exclusions python/cudf/pyproject.toml, python/dask_cudf/pyproject.toml, python/pylibcudf/pyproject.toml	Core and optional dependency entries across three project configurations update cupy-cuda13x>=13.6.0,!=14.0.0 to cupy-cuda13x>=13.6.0,!=14.0.0,!=14.1.0.
Central dependency manifest CuPy 14.1.0 exclusions dependencies.yaml	Three sections expand CuPy version constraints to exclude version 14.1.0: test requirements for cupy-cuda12x and cupy-cuda13x in test_python_cudf_common, conda list for depends_on_cupy, and requirements/pyproject matrices for CUDA-specific variants.
Polars GPU engine benchmark documentation docs/cudf/source/cudf_polars/index.md	Benchmark section replaces static figures with new narrative text and two rendered table-based figure blocks for scale factors 1K and 3K, removes CUDA managed memory and reproduction details, and updates reference to PDS-DS queries.

🎯 2 (Simple) | ⏱️ ~10 minutes

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}