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
45 lines
1.8 KiB
ReStructuredText
45 lines
1.8 KiB
ReStructuredText
.. _libcudacxx-setup-building:
|
|
|
|
Building & Testing libcu++
|
|
==========================
|
|
|
|
libcu++ can be build and tested as shown in our `contributor guidelines <https://github.com/NVIDIA/cccl/blob/main/CONTRIBUTING.md#building-and-testing>`_.
|
|
|
|
However, often only a small subset of the full test suite needs to be run during development. For that we rely on ``lit``.
|
|
After libcu++ has been configured either through the build scripts or directly via a cmake preset one can then run.
|
|
|
|
.. code:: bash
|
|
|
|
cd build
|
|
lit libcudacxx/RELATIVE_PATH_TO_TEST_OR_SUBFOLDER -sv
|
|
|
|
This will build and run all tests within ``RELATIVE_PATH_TO_TEST_OR_SUBFOLDER`` which must be a valid path within the CCCL.
|
|
Note that the name of the top level folder is the same as the name of the preset. For the build script the default is
|
|
``libcudacxx``. As an example this is how to run all tests for ``cuda::std::span``, which are located in
|
|
``libcudacxx/test/libcudacxx/std/containers/views/views.span``
|
|
|
|
.. code:: bash
|
|
|
|
cd build
|
|
|
|
# Builds all tests within libcudacxx/test/libcudacxx/std/containers/views/views.span
|
|
lit libcudacxx/libcudacxx/test/libcudacxx/std/containers/views/views.span -sv
|
|
|
|
# Builds the individual test array.pass.cpp
|
|
lit libcudacxx/libcudacxx/test/libcudacxx/std/containers/views/views.span/span.cons/array.pass.cpp -sv
|
|
|
|
If only building the tests and not running them is desired one can pass ``-Dexecutor="NoopExecutor()"`` to the lit invocation.
|
|
This is especially useful if the machine has no GPU or testing a different architecture
|
|
|
|
.. code:: bash
|
|
|
|
cd build
|
|
lit libcudacxx/RELATIVE_PATH_TO_TEST_OR_SUBFOLDER -sv -Dexecutor="NoopExecutor()"
|
|
|
|
Finally different standard modes can be tested by passing e.g ``--param=std=c++20``
|
|
|
|
.. code:: bash
|
|
|
|
cd build
|
|
lit libcudacxx/RELATIVE_PATH_TO_TEST_OR_SUBFOLDER -sv --param=std=c++20
|