[KyleFromNVIDIA]

Description

Use CCCL buffers instead of RMM vectors where possible in inline implementations in the headers.

Issue: rapidsai/build-planning#321

Checklist

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

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: NVIDIA/cudf
• Branch: main
• File: .coderabbit.yaml
• Commit: f5cb11b
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Enterprise
• Run ID: 623f521f-4cdb-4e2f-825d-8d37ee19b0fb

Commits

Reviewing files that changed from the base of the PR and between 6fea1f9 and 87da2ca.

Files selected for processing: 2

• cpp/include/cudf/reduction/detail/approx_distinct_count.cuh
• cpp/src/reductions/approx_distinct_count.cu

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

Summary

Summary by CodeRabbit

• Refactor
  • Updated internal temporary storage across data gathering, reductions, string operations, and related workflows. Existing calculation, comparison, and output behavior remains unchanged.
  • No user-facing functionality changes are included in this update.

Internal storage and temporary allocations across detail, reduction, and strings code now use cuda::device_buffer instead of RMM device vectors. Approximate distinct-count sketch storage and its span-backed constructor were also updated.

Changes: Device buffer migration

Layer / File(s)	Summary
Detail storage and temporary allocations cpp/include/cudf/detail/device_scalar.hpp, cpp/include/cudf/detail/gather.cuh, cpp/include/cudf/detail/labeling/label_segments.cuh, cpp/include/cudf/detail/null_mask.cuh, cpp/include/cudf/detail/row_operator/primitive_lexicographic.cuh, cpp/include/cudf/detail/utilities/batched_memset.hpp	Device scalar storage and temporary allocations for gather, labeling, null masks, lexicographic comparison, and batched memset now use CUDA device buffers.
Reduction storage cpp/include/cudf/reduction/detail/approx_distinct_count.cuh, cpp/include/cudf/reduction/detail/segmented_reduction.cuh, cpp/src/reductions/approx_distinct_count.cu	Reduction-owned storage uses CUDA device buffers. The approximate distinct-count sketch visitors recognize the new owning storage type, and the span-backed constructor no longer accepts or stores a memory resource.
Strings temporary allocations cpp/include/cudf/strings/detail/copy_if_else.cuh, cpp/include/cudf/strings/detail/strings_children.cuh, cpp/include/cudf/strings/detail/strings_column_factories.cuh	Temporary indices, sizes, character sizes, and valid counts use CUDA device buffers.

Estimated code review effort: 2 (Simple) | ~12 minutes

Change: Refactor

Suggested reviewers: pointkernel

Merge Risk: Minimal · up to 87da2

The migration introduces no confirmed merge-blocking risk and is suitable to merge with normal checks.

Pre-merge checks: 4 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentat…	Write docstrings for the functions missing them to satisfy the coverage threshold.

Passed checks: 4 passed

Check name	Status	Explanation
Title check		The title clearly and concisely describes the main change: replacing scratch rmm::device_uvector storage with cuda::device_buffer in headers.
Description check		The description directly relates to the changeset and states the objective of using CCCL buffers instead of RMM vectors in inline header implementations.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 2 functions across 3 files. (1 skipped: 1 unsupported.)

Resolution

Write docstrings for the functions missing them to satisfy the coverage threshold.

Finishing Touches

Generate unit tests (beta)

• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

[coderabbitai]

• Fix comments: Fix CodeRabbit comments on this PR

[bdice]

Approving but there's an opportunity for cleanup, I think.

[KyleFromNVIDIA]

/merge