Drop usage of thrust iterators

[miscco]

We can use the more modern cuda iterators instead

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• Refactor
  • Replaced internal iterator/iterator-utility implementations with CUDA-provided equivalents. No changes to public APIs or runtime behavior.
• Chores
  • Updated copyright years and removed unused internal includes.

Walkthrough

Replaced Thrust iterator utilities with CUDA C++ iterator equivalents across benchmarks, headers, and tests; updated iterator type aliases in device_uvector to cuda::std::reverse_iterator; removed several rmm/detail/thrust_namespace.h includes and bumped copyright years.

Changes

Cohort / File(s)	Summary
Benchmarks & Replay cpp/benchmarks/replay/replay.cpp	Replaced thrust::make_constant_iterator / thrust::make_discard_iterator with cuda::make_constant_iterator / cuda::make_discard_iterator; added <cuda/iterator> include and removed Thrust iterator includes; copyright year updated.
Device uvector iterators cpp/include/rmm/device_uvector.hpp	Switched reverse_iterator and const_reverse_iterator aliases from thrust::reverse_iterator<...> to cuda::std::reverse_iterator<...> and added the corresponding CUDA std iterator include. (Public type-alias resolution changed.)
Memory resource iterators cpp/include/rmm/mr/fixed_size_memory_resource.hpp, cpp/include/rmm/mr/pool_memory_resource.hpp	Replaced Thrust counting/transform iterator usage with cuda::make_counting_iterator / cuda::make_transform_iterator and <cuda/iterator> include; removed Thrust iterator headers and rmm/detail/thrust_namespace.h include.
Tests (logger) cpp/tests/logger_tests.cpp	Replaced thrust::make_zip_iterator and Thrust zip iterator include with cuda::make_zip_iterator and <cuda/iterator>; copyright year updated.
Async memory resource headers cpp/include/rmm/mr/cuda_async_memory_resource.hpp, cpp/include/rmm/mr/cuda_async_managed_memory_resource.hpp, cpp/include/rmm/mr/cuda_async_view_memory_resource.hpp	Removed #include <rmm/detail/thrust_namespace.h> from headers; no other API or behavioral changes.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

🚥 Pre-merge checks | ✅ 3 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Merge Conflict Detection	⚠️ Warning	❌ Merge conflicts detected (9 files): ⚔️ cpp/benchmarks/replay/replay.cpp (content) ⚔️ cpp/include/rmm/device_uvector.hpp (content) ⚔️ cpp/include/rmm/mr/cuda_async_managed_memory_resource.hpp (content) ⚔️ cpp/include/rmm/mr/cuda_async_memory_resource.hpp (content) ⚔️ cpp/include/rmm/mr/cuda_async_view_memory_resource.hpp (content) ⚔️ cpp/include/rmm/mr/detail/stream_ordered_memory_resource.hpp (content) ⚔️ cpp/include/rmm/mr/fixed_size_memory_resource.hpp (content) ⚔️ cpp/include/rmm/mr/pool_memory_resource.hpp (content) ⚔️ cpp/tests/logger_tests.cpp (content) These conflicts must be resolved before merging into main.	Resolve conflicts locally and push changes to this branch.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Drop usage of thrust iterators' directly and clearly describes the main objective of the changeset, which systematically replaces thrust iterators with CUDA equivalents across multiple files.
Description check	✅ Passed	The description 'We can use the more modern cuda iterators instead' is directly related to the changeset, explaining the rationale for replacing thrust iterators with CUDA iterators.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

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

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

⚔️ Resolve merge conflicts (beta)

• Auto-commit resolved conflicts to branch drop_thrust_iterator
• Post resolved changes as copyable diffs in a comment

---

No actionable comments were generated in the recent review. 🎉

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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

[coderabbitai]

Actionable comments posted: 2

🤖 Fix all issues with AI agents

In `@cpp/include/rmm/mr/fixed_size_memory_resource.hpp`:
- Line 17: Remove the now-unused include of <rmm/detail/thrust_namespace.h> from
cpp/include/rmm/mr/fixed_size_memory_resource.hpp (it was used previously for
Thrust iterator aliases but is no longer needed after switching to
<cuda/iterator> and cuda::make_transform_iterator /
cuda::make_counting_iterator); locate the `#include`
"<rmm/detail/thrust_namespace.h>" and delete that line and then verify no other
references to symbols from rmm::detail::thrust_namespace remain in the file
(e.g., usages around FixedSizeMemoryResource or related helper functions).

In `@cpp/tests/logger_tests.cpp`:
- Around line 99-102: The tests use the non-existent helper
cuda::make_zip_iterator; replace its usage by directly constructing
cuda::zip_iterator for the begin and end iterators so the compiler can find the
correct API. Specifically, update the creation of begin and end that currently
reference cuda::make_zip_iterator to construct cuda::zip_iterator objects using
expected_events.begin()/end() and actual_events.begin()/end() respectively (the
variables named begin and end and the containers expected_events and
actual_events). Ensure both begin and end use the direct cuda::zip_iterator
constructor syntax so the code compiles.

[bdice]

@miscco If you don’t mind, please look at the CodeRabbit suggestions and comment on whether they are useful/true/false. If there are incorrect comments, I am interested in the “Analysis Chain” and where it is right or wrong. I’m refining the AI review prompts and I might need to give it more context on how to consult against CCCL.

[bdice]

/merge