[Backport to 3.4] Remove most _CCCL_GRID_CONSTANT annotations, including merge sort

[bernhardmgruber]

This PR backports the following two PRs:

• Remove most CCCL_GRID_CONSTANT annotations #9858 (cherry-picked from d8fd76b)
• Remove _CCCL_GRID_CONSTANT from merge sort kernel parameters #9829 (cherry-picked from aeaa386)

[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: edbc1696-0285-4862-a3e4-dbaa75d9b719

📥 Commits

Reviewing files that changed from the base of the PR and between bf8e9d2 and 314504a.

📒 Files selected for processing (21)

• cub/cub/device/dispatch/dispatch_adjacent_difference.cuh
• cub/cub/device/dispatch/dispatch_batch_memcpy.cuh
• cub/cub/device/dispatch/dispatch_merge.cuh
• cub/cub/device/dispatch/dispatch_reduce_by_key.cuh
• cub/cub/device/dispatch/dispatch_rle.cuh
• cub/cub/device/dispatch/dispatch_scan_by_key.cuh
• cub/cub/device/dispatch/dispatch_segmented_sort.cuh
• cub/cub/device/dispatch/dispatch_topk.cuh
• cub/cub/device/dispatch/kernels/kernel_for_each.cuh
• cub/cub/device/dispatch/kernels/kernel_histogram.cuh
• cub/cub/device/dispatch/kernels/kernel_merge_sort.cuh
• cub/cub/device/dispatch/kernels/kernel_radix_sort.cuh
• cub/cub/device/dispatch/kernels/kernel_reduce.cuh
• cub/cub/device/dispatch/kernels/kernel_scan.cuh
• cub/cub/device/dispatch/kernels/kernel_segmented_radix_sort.cuh
• cub/cub/device/dispatch/kernels/kernel_segmented_reduce.cuh
• cub/cub/device/dispatch/kernels/kernel_segmented_scan.cuh
• cub/cub/device/dispatch/kernels/kernel_segmented_sort.cuh
• cub/cub/device/dispatch/kernels/kernel_three_way_partition.cuh
• cub/cub/device/dispatch/kernels/kernel_transform.cuh
• cub/cub/device/dispatch/kernels/kernel_unique_by_key.cuh

---

📝 Walkthrough

Summary by CodeRabbit

• Refactor
  • Updated internal GPU kernel interfaces across sorting, scanning, reduction, histogram, transform, and partition operations.
  • Simplified parameter declarations while preserving existing processing logic and runtime behavior.
  • No user-facing functionality or API behavior changes are expected.

Walkthrough

Changes

The PR removes _CCCL_GRID_CONSTANT from CUDA kernel parameter declarations across CUB device dispatch, sorting, reduction, scan, partition, histogram, transform, and related kernels. Existing parameter types, const semantics, call sites, and kernel bodies remain unchanged.

CUB kernel signature updates

Layer / File(s)	Summary
Dispatch entry-point signatures cub/cub/device/dispatch/dispatch_*.cuh	Adjacent-difference, batch-memcpy, merge, reduce-by-key, RLE, scan-by-key, segmented-sort, and TopK kernel parameters no longer use _CCCL_GRID_CONSTANT.
General algorithm kernel signatures cub/cub/device/dispatch/kernels/kernel_for_each.cuh, kernel_histogram.cuh, kernel_reduce.cuh, kernel_scan.cuh, kernel_transform.cuh, kernel_unique_by_key.cuh	For-each, histogram, reduction, scan, transform, and unique-by-key kernel declarations use plain parameter qualifiers.
Sort kernel signatures cub/cub/device/dispatch/kernels/kernel_merge_sort.cuh, kernel_radix_sort.cuh	Merge-sort and radix-sort stage kernel declarations remove _CCCL_GRID_CONSTANT without changing kernel logic.
Segmented algorithm signatures cub/cub/device/dispatch/kernels/kernel_segmented_*.cuh	Segmented radix sort, reduction, scan, and sort kernels retain their existing types while removing the grid-constant qualifier.
Partition kernel signatures cub/cub/device/dispatch/kernels/kernel_three_way_partition.cuh	Three-way partition kernel parameters remove _CCCL_GRID_CONSTANT from iterator, output, scalar, and streaming-context arguments.

Possibly related PRs

• NVIDIA/cccl#9858: Updates the same CUB dispatch kernel parameter declarations by removing _CCCL_GRID_CONSTANT.

Suggested labels: backport branch/3.4.x

Suggested reviewers: davebayer, gevtushenko

---

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

[github-actions]

🥳 CI Workflow Results

🟩 Finished in 2h 24m: Pass: 100%/283 | Total: 11d 22h | Max: 2h 24m | Hits: 15%/1080970

See results here.