Use static cudart

[KyleFromNVIDIA]

Issue: rapidsai/build-planning#235

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: bd10c6d9-dca4-447d-bf1b-7372d6c6a010

📥 Commits

Reviewing files that changed from the base of the PR and between 3c13543 and 7fd415f.

📒 Files selected for processing (1)

• cpp/CMakeLists.txt

🚧 Files skipped from review as they are similar to previous changes (1)

• cpp/CMakeLists.txt

---

📝 Walkthrough

Summary by CodeRabbit

• Chores

  • Enforced a consistent static CUDA runtime across builds for more predictable packaging.
  • Removed a redundant CUDA runtime host dependency from packaged outputs to simplify installs.

• Documentation

  • Updated build docs to reflect the revised CUDA/CMake configuration and resulting packaging behavior.

Walkthrough

Removes the CUDA_STATIC_RUNTIME CMake option, hardcodes static CUDA runtime initialization (USE_STATIC ON), updates nvforest++ linking to use CUDA::cudart_static, removes cuda-cudart from conda host requirements, and deletes the option from build docs.

Changes

CUDA static runtime hardening

Layer / File(s)	Summary
CMake CUDA static runtime hardcoding cpp/CMakeLists.txt, python/libnvforest/CMakeLists.txt	Removes the CUDA_STATIC_RUNTIME option and its verbose status; rapids_cuda_init_runtime is called with USE_STATIC ON. nvforest++ now includes CUDA::cudart_static in its public link list. The python CMakeLists no longer sets CUDA_STATIC_RUNTIME before adding the cpp subdirectory.
Conda recipe dependency alignment conda/recipes/libnvforest/recipe.yaml	Removes the cuda-cudart host requirement from both libnvforest and libnvforest-tests outputs.
Build documentation updates docs/source/build.rst	Removes the CUDA_STATIC_RUNTIME flag row from the libnvforest++ CMake options table.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• rapidsai/nvforest#89: Modifies the same conda recipe and CMake CUDA static-runtime handling files with related configuration changes.
• rapidsai/nvforest#87: Adjusts CUDA-related dependencies in the conda recipe (removes CUDA math library dependencies), related to dependency alignment.

Suggested reviewers

• csadorf
• hcho3
• gforsyth

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Use static cudart' directly reflects the main change: unconditionally linking against the static CUDA runtime library across all build configurations.
Description check	✅ Passed	The description references a relevant issue that motivated the changes (build-planning issue #235), which is appropriate context for the PR's static CUDA runtime linking updates.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@cpp/CMakeLists.txt`:
- Around line 297-301: The current INTERFACE $<COMPILE_ONLY:CUDA::cudart_static>
on target_link_libraries for ${NVFOREST_CPP_TARGET} only propagates compile info
and will break static consumers at link time; change that generator expression
to a conditional one that applies LINK_ONLY:CUDA::cudart_static when the
consuming target is a static library and omits the CUDA runtime for shared
builds (i.e., replace the COMPILE_ONLY use with a conditional generator
expression that yields $<LINK_ONLY:CUDA::cudart_static> for static-type
consumers of ${NVFOREST_CPP_TARGET} and nothing for shared).

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 7dac64f1-41cb-4fad-b406-4bfd5cae8bb5

📥 Commits

Reviewing files that changed from the base of the PR and between d35f3fd and 3c13543.

📒 Files selected for processing (1)

• cpp/CMakeLists.txt

[KyleFromNVIDIA]

/merge