[libcu++] Implement P3060R3 Add `std::views::indices(n)` by davebayer · Pull Request #10700 · NVIDIA/cccl

davebayer · 2026-08-07T06:11:53Z

coderabbitai · 2026-08-07T06:15:15Z

📝 Walkthrough

Summary by CodeRabbit

New Features
- Added cuda::std::views::indices for creating integer-index ranges from zero to a specified size.
- Supports integer-like inputs, compile-time evaluation, range indexing, and size queries.
Bug Fixes
- Improved range size calculation for signed starting values, including correct handling of negative values.
Compatibility
- Added a feature-test indicator for detecting views::indices support.

Walkthrough

The PR adds cuda::std::views::indices, updates iota_view::size() for unsigned starts, defines __cccl_lib_ranges_indices, and adds compile-time and runtime tests.

Changes

views::indices

Layer / File(s)	Summary
Implement views::indices and iota_view sizing `libcudacxx/include/cuda/std/__ranges/iota_view.h`	The implementation adds required headers, limits negative-value handling to signed starts, and constructs an `iota_view` from zero to an integer-like size.
Publish and test the indices contract `libcudacxx/include/cuda/std/version`, `libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp`	The version header defines `__cccl_lib_ranges_indices` as `202506L`. Tests cover constraints, values, size, array-size construction, runtime execution, and constexpr execution.

Assessment against linked issues

Objective	Addressed	Explanation
Implement `std::views::indices(n)` from P3060R3 [`#10356`]	✅

Suggested reviewers: griwes

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

coderabbitai

🧹 Nitpick comments (2)

libcudacxx/include/cuda/std/__ranges/iota_view.h (1)

270-270: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

suggestion: Declare __size as const _Tp __size. Line 270 does not modify this parameter. As per coding guidelines, “All variables that are not modified must be declared const.”

Source: Coding guidelines

libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp (1)

27-57: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

suggestion: Rename HasIndices, HasIndicesFn, NotIntegerLike, and test_SFINAE to snake_case. These are not template parameters or CUB public API symbols. As per coding guidelines, “Use snake_case for all other symbols.”

Source: Coding guidelines

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: dd60016d-4852-4385-8a66-ec1cda5adf54

📥 Commits

Reviewing files that changed from the base of the PR and between 25e6f45 and 237e92b.

📒 Files selected for processing (3)

libcudacxx/include/cuda/std/__ranges/iota_view.h
libcudacxx/include/cuda/std/version
libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp

coderabbitai · 2026-08-07T06:15:23Z

Note

GitHub couldn't provide a complete incremental comparison for this pull request, so CodeRabbit is performing a full review instead. This review may take a little longer.

coderabbitai

🧹 Nitpick comments (1)

libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp (1)

27-46: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

suggestion: Rename HasIndices, HasIndicesFn, NotIntegerLike, and test_SFINAE to snake_case. Keep SizeType and T in PascalCase because they are template parameters. As per coding guidelines, “Use snake_case for all other symbols, except that the CUB public API uses PascalCase.”

Source: Coding guidelines

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 9089676e-6bef-4dbb-8a8d-567e84e0a086

📥 Commits

Reviewing files that changed from the base of the PR and between 25e6f45 and 73ea3a2.

📒 Files selected for processing (3)

libcudacxx/include/cuda/std/__ranges/iota_view.h
libcudacxx/include/cuda/std/version
libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp

🚧 Files skipped from review as they are similar to previous changes (2)

libcudacxx/include/cuda/std/__ranges/iota_view.h
libcudacxx/include/cuda/std/version

github-actions · 2026-08-07T06:48:35Z

⏱️ CCCL compile-time benchmark comparison: Public headers compile-time bench

Result: 0 regression row(s), 11 improvement row(s) above threshold.

Run	Value
Config	`public-headers-gcc13`
Baseline	`origin/main`
Preset	`all-dev`
Targets	`cub.headers.base, thrust.cpp.cuda.headers.base, libcudacxx.test.public_headers`
GPU / launch args	`rtx2080` / `--cuda 13.3 --host gcc13`

Artifacts: reports and traces

TU total compilation

-f total-compilation inclusive --sort total

🟢 TU total compilation — Improvements

Rank	Improvement impact	Selected Δ	Baseline	Current	Event	Matched traces
1	`3.044159`	`-3.044159`	`13.488601`	`10.444442`	Total Compilation Time: `cub/device/device_segmented_sort.cuh`	1

Direct file processing

-f file-processing exclusive --sort total

🟢 Direct file processing — Improvements

Rank	Improvement impact	Selected Δ	Baseline	Current	Event	Matched traces
1	`1.178625`	`-1.178625`	`9.630218`	`8.451593`	Processing Header File: `libcudacxx/include/cuda/std/__cccl/prologue.h`	550
2	`1.004304`	`-1.004304`	`20.988506`	`19.984202`	Processing Header File: `libcudacxx/include/cuda/__device/physical_device.h`	94
3	`0.478540`	`-0.478540`	`3.417435`	`2.938895`	Processing Header File: `libcudacxx/include/cuda/std/__cccl/epilogue.h`	550
4	`0.435249`	`-0.435249`	`11.536714`	`11.101465`	Processing Header File: `libcudacxx/include/cuda/std/__type_traits/type_list.h`	477
5	`0.388171`	`-0.388171`	`11.095746`	`10.707575`	Processing Header File: `libcudacxx/include/cuda/std/__iterator/concepts.h`	437
6	`0.326840`	`-0.326840`	`6.229877`	`5.903037`	Processing Header File: `libcudacxx/include/cuda/std/__tuple_dir/vector_types.h`	238
7	`0.258641`	`-0.258641`	`4.205050`	`3.946409`	Processing Header File: `libcudacxx/include/cuda/std/__utility/pair.h`	404
8	`0.256344`	`-0.256344`	`5.771235`	`5.514891`	Processing Header File: `libcudacxx/include/cuda/std/__iterator/iterator_traits.h`	464
9	`0.229500`	`-0.229500`	`3.791927`	`3.562427`	Processing Header File: `libcudacxx/include/cuda/std/__tuple_dir/tuple.h`	243
10	`0.203049`	`-0.203049`	`7.792895`	`7.589846`	Processing Header File: `cub/cub/util_type.cuh`	279

github-actions · 2026-08-07T10:42:59Z

😬 CI Workflow Results

🟥 Finished in 4h 27m: Pass: 88%/115 | Total: 4d 09h | Max: 4h 27m | Hits: 43%/1225052

See results here.

AI failure analysis

1. Default-initialized SomeInt array breaks NVCC/MSVC constexpr evaluation · 13 jobs

Explanation: The new `indices.pass.cpp` test fails during `static_assert(test())` in every NVCC/MSVC matrix entry. The compiler cannot treat the synthesized default-constructor closure for the test's uninitialized `cuda::std::array<SomeInt, 5>` as constexpr, preventing all 13 Windows builds from completing.

Evidence:

libcu++ nvcc MSVC / [CTK12.0 MSVC14.39 C++20] Build(amd64), step 4

C:\cccl\libcudacxx\test\libcudacxx\std\ranges\range.factories\range.iota.view\indices.pass.cpp(110): error C2131: expression did not evaluate to a constant
C:/cccl/libcudacxx/include\cuda/std/array(224): note: failure was caused by call of undefined function or one not declared 'constexpr'
C:/cccl/libcudacxx/include\cuda/std/array(224): note: see usage of 'SomeInt::__dflt_ctor_closure'

Root cause: The PR declares `cuda::std::array<SomeInt, 5> v;` inside a function evaluated by `static_assert`. Under every tested NVCC/MSVC combination, that default initialization produces a synthesized `SomeInt::__dflt_ctor_closure` that the compiler cannot evaluate as constexpr; the neighboring integer array is explicitly value-initialized with `{}` and does not fail. Sources: libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp:90, libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp:110, libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/types.h:20.

Suggested next steps: Change the declaration at line 90 to `cuda::std::array<SomeInt, 5> v{};`, then run the focused `indices.pass.cpp` libcudacxx lit test with NVCC/MSVC in C++17 and C++20. Verify at least one oldest and one newest failing toolchain before rerunning the Windows matrix.

Copy this prompt into a coding agent

Repository: https://github.com/NVIDIA/cccl
Workflow run: https://github.com/NVIDIA/cccl/actions/runs/31153186384
Failure group: Default-initialized SomeInt array breaks NVCC/MSVC constexpr evaluation
Affected jobs:
- libcu++ nvcc MSVC / [CTK12.0 MSVC14.39 C++20] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164850
- libcu++ nvcc MSVC / [CTK12.9 MSVC14.44 C++20] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164885
- libcu++ nvcc MSVC / [CTK13.3 MSVC14.50 C++20] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164886
- libcu++ nvcc MSVC / [CTK12.0 MSVC14.39 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164889
- libcu++ nvcc MSVC / [CTK13.3 MSVC14.29 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164914
- libcu++ nvcc MSVC / [CTK12.9 MSVC14.29 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164922
- libcu++ nvcc MSVC / [CTK12.9 MSVC14.44 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164935
- libcu++ nvcc MSVC / [CTK13.3 MSVC14.44 C++20] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164936
- libcu++ nvcc MSVC / [CTK13.3 MSVC14.50 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164956
- libcu++ nvcc MSVC / [CTK13.0 MSVC14.44 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164976
- libcu++ nvcc MSVC / [CTK13.0 MSVC14.29 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164982
- libcu++ nvcc MSVC / [CTK13.0 MSVC14.44 C++20] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164983
- libcu++ nvcc MSVC / [CTK12.0 MSVC14.29 C++17] Build(amd64): https://github.com/NVIDIA/cccl/actions/runs/31153186384/job/92787164998

Reproduce the NVCC/MSVC failure narrowly for `libcudacxx/test/libcudacxx/std/ranges/range.factories/range.iota.view/indices.pass.cpp`. The failure occurs at `static_assert(test())` because line 90 declares `cuda::std::array<SomeInt, 5> v;`, and NVCC/MSVC reports that `SomeInt::__dflt_ctor_closure` is not constexpr. Implement the minimal fix by changing that declaration to `cuda::std::array<SomeInt, 5> v{};`, matching the value-initialized integer-array test above it. Run the focused lit test in C++17 and C++20 with representative NVCC/MSVC configurations, including an older and newer toolkit/compiler pair; only investigate the `views::indices` implementation if the value-initialization fix does not resolve the constexpr failure.

Jobs:

davebayer requested a review from a team as a code owner August 7, 2026 06:11

davebayer requested a review from griwes August 7, 2026 06:11

github-project-automation Bot added this to CCCL Aug 7, 2026

github-project-automation Bot moved this to Todo in CCCL Aug 7, 2026

cccl-authenticator-app Bot moved this from Todo to In Review in CCCL Aug 7, 2026

[libcu++] Implement P3060R3 Add std::views::indices(n)

73ea3a2

davebayer force-pushed the views_indices branch from 237e92b to 73ea3a2 Compare August 7, 2026 06:13

coderabbitai Bot reviewed Aug 7, 2026

View reviewed changes

miscco approved these changes Aug 7, 2026

View reviewed changes

davebayer enabled auto-merge (squash) August 7, 2026 07:55

Navigation Menu

Uh oh!

[libcu++] Implement P3060R3 Add `std::views::indices(n)` - #10700

[libcu++] Implement P3060R3 Add `std::views::indices(n)`#10700
davebayer wants to merge 1 commit into
NVIDIA:mainfrom
davebayer:views_indices

davebayer commented Aug 7, 2026

Uh oh!

coderabbitai Bot commented Aug 7, 2026 •

edited

Loading

Summary by CodeRabbit

Walkthrough

Changes

Assessment against linked issues

Uh oh!

coderabbitai Bot left a comment

Uh oh!

coderabbitai Bot commented Aug 7, 2026

Uh oh!

coderabbitai Bot left a comment

Uh oh!

github-actions Bot commented Aug 7, 2026

Uh oh!

github-actions Bot commented Aug 7, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

davebayer commented Aug 7, 2026

Uh oh!

coderabbitai Bot commented Aug 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary by CodeRabbit

Walkthrough

Changes

Assessment against linked issues

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

coderabbitai Bot commented Aug 7, 2026

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Aug 7, 2026

⏱️ CCCL compile-time benchmark comparison: Public headers compile-time bench

TU total compilation

Direct file processing

Uh oh!

github-actions Bot commented Aug 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

😬 CI Workflow Results

🟥 Finished in 4h 27m: Pass: 88%/115 | Total: 4d 09h | Max: 4h 27m | Hits: 43%/1225052

AI failure analysis

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

coderabbitai Bot commented Aug 7, 2026 •

edited

Loading

github-actions Bot commented Aug 7, 2026 •

edited

Loading