2a7ca101d7
Sparse-checkout from NVIDIA/cccl main branch to complete cccl_upstream: Added: - python/cuda_cccl/ (226 files) — Python bindings for device-level algorithms Critical for muh toolchain: cuda.compute.reduce_into, scan, radix_sort, etc. Includes 204 .py files with full test coverage for all 27 algorithms - ci/ (163 files) — Build/test infrastructure build_cub.sh, test_cub.sh, build_and_test_targets.sh, matrix.yaml Directly maps to our [INFRA-CI] and [INFRA-BUILD] items - .agent/skills/ (7 files) — NVIDIA's own agent skills for CCCL cccl-style/SKILL.md, cccl-test/SKILL.md, sass-diff/SKILL.md - docs/ (491 files) — Official CCCL documentation CI references, CMake guides, Python compute docs, libcudacxx PTX docs - test/ (12 files) — Top-level integration tests (cuda_smoke, stdpar) - Root configs: .clang-format, .clang-tidy, CONTRIBUTING.md, pyproject.toml - CLAUDE.md symlink → AGENTS.md (NVIDIA's standard) cccl_upstream now mirrors full NVIDIA/cccl structure: Before: 42M (cub + thrust + libcudacxx + cudax + c + examples + benchmarks) After: 53M (+python +ci +docs +.agent +test +configs) This completes the CCCL base needed for: - [muh-bench] items: ci/util/build_and_test_targets.sh for targeted builds - [CCCL-verify] items: python/cuda_cccl/tests/ as reference implementations - [CCCL-test] items: ci/test_cub.sh, ci/test_thrust.sh - Agent workflow: .agent/skills/ for consistent style and test patterns
65 lines
1.6 KiB
ReStructuredText
65 lines
1.6 KiB
ReStructuredText
.. _libcudacxx-extended-api-thread-groups:
|
|
|
|
Thread Groups
|
|
=============
|
|
|
|
.. code:: cuda
|
|
|
|
struct ThreadGroup {
|
|
static constexpr cuda::thread_scope thread_scope;
|
|
Integral size() const;
|
|
Integral thread_rank() const;
|
|
void sync() const;
|
|
};
|
|
|
|
The *ThreadGroup concept* defines the requirements of a type that represents a group of cooperating threads.
|
|
|
|
The `CUDA Cooperative Groups Library <https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#group-collectives>`_
|
|
provides a number of types that satisfy this concept.
|
|
|
|
Data Members
|
|
------------
|
|
|
|
.. list-table::
|
|
:widths: 25 75
|
|
:header-rows: 0
|
|
|
|
* - ``thread_scope``
|
|
- The scope at which ``ThreadGroup::sync()`` synchronizes memory operations and thread execution.
|
|
|
|
Member Functions
|
|
----------------
|
|
|
|
.. list-table::
|
|
:widths: 25 75
|
|
:header-rows: 0
|
|
|
|
* - ``size``
|
|
- Returns the number of participating threads.
|
|
* - ``thread_rank``
|
|
- Returns a unique value for each participating thread (``0 <= ThreadGroup::thread_rank() < ThreadGroup::size()``).
|
|
* - ``sync``
|
|
- Synchronizes the participating threads.
|
|
|
|
Notes
|
|
-----
|
|
|
|
This concept is defined for documentation purposes but is not materialized in the library.
|
|
|
|
Example
|
|
-------
|
|
|
|
.. code:: cuda
|
|
|
|
#include <cuda/atomic>
|
|
#include <cuda/std/cstddef>
|
|
|
|
struct single_thread_group {
|
|
static constexpr cuda::thread_scope thread_scope = cuda::thread_scope::thread_scope_thread;
|
|
cuda::std::size_t size() const { return 1; }
|
|
cuda::std::size_t thread_rank() const { return 0; }
|
|
void sync() const {}
|
|
};
|
|
|
|
`See it on Godbolt <https://godbolt.org/z/6c16KxqY7>`_
|