feat(ascend): add top-k top-p sampler

[zhangyue207]

Summary

• Add the public TopKTopPSampler base API in src/base/top_k_top_p_sampler.h.
• Add an Ascend ACLNN implementation in src/native/ascend/ops/top_k_top_p_sampler/kernel.h.
• Add focused Ascend tests in tests/test_top_k_top_p_sampler.py.

Motivation

Closes #

This adds a logits-level Ascend sampler boundary needed by inference pipelines that already prepare top-k and top-p controls outside InfiniOps.

Type of Change

• feat — new feature / new operator / new platform
• fix — bug fix
• perf — performance improvement (no behavioral change)
• refactor — code restructuring without behavior change
• test — adding or fixing tests only
• docs — documentation only
• build / ci — build system or CI configuration
• chore — tooling, formatting, or other non-code changes
• Breaking change (requires a ! in the Conventional Commits prefix or a BREAKING CHANGE: footer)

Platforms Affected

• CPU (WITH_CPU)
• NVIDIA (WITH_NVIDIA)
• Iluvatar (WITH_ILUVATAR)
• MetaX (WITH_METAX)
• Cambricon (WITH_CAMBRICON)
• Moore (WITH_MOORE)
• Ascend (WITH_ASCEND)
• PyTorch C++ bindings (WITH_TORCH)
• Build system / CMake / CI
• Python bindings / user-facing API

Test Results on Supported Platforms

Platform	Built	pytest Result	Notes / Hardware
NVIDIA	N/A	N/A	Not touched.
Iluvatar	N/A	N/A	Not touched.
MetaX	N/A	N/A	Not touched.
Cambricon	N/A	N/A	Not touched.
Moore	N/A	N/A	Not touched.
Ascend	Yes	6 passed in 1.81s	Project Ascend test container.

Full `pytest` output (optional)

python3 -m pip install .[dev] --no-build-isolation -C cmake.define.WITH_ASCEND=ON -C cmake.define.BUILD_CUSTOM_KERNEL=OFF -C cmake.define.AUTO_DETECT_DEVICES=OFF -C cmake.define.GENERATE_PYTHON_BINDINGS=ON
Successfully built InfiniOps
Successfully installed InfiniOps-0.1.0

ASCEND_VISIBLE_DEVICES=0 ASCEND_RT_VISIBLE_DEVICES=0 pytest tests/test_top_k_top_p_sampler.py --devices ascend -v --tb=short
collected 6 items
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler[npu-0-dtype0-shape0] PASSED [ 16%]
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler[npu-0-dtype0-shape1] PASSED [ 33%]
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler[npu-0-dtype1-shape0] PASSED [ 50%]
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler[npu-0-dtype1-shape1] PASSED [ 66%]
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler_optional_p[npu-0-dtype0] PASSED [ 83%]
tests/test_top_k_top_p_sampler.py::test_top_k_top_p_sampler_optional_p[npu-0-dtype1] PASSED [100%]
============================== 6 passed in 1.81s ===============================

Benchmark / Performance Impact

N/A. This PR adds a correctness-tested operator wrapper and does not include a performance benchmark.

Notes for Reviewers

API alignment note:

• Reference checked: vLLM vllm.v1.sample.ops.topk_topp_sampler.TopKTopPSampler, whose native path receives logits, generators, k, and p, applies top-k/top-p filtering, and samples from logits.
• InfiniOps base signature: TopKTopPSampler(logits, k, p, out) and operator()(logits, k, p, out).
• Intentional deviations: InfiniOps uses an output tensor instead of returning sampled ids, does not expose Python generators, and does not return processed logits/logprobs. Temperature scaling remains caller-owned.
• Current Ascend ACLNN implementation is verified for the deterministic top_k == 1 subset; broader stochastic sampling is left out of this minimal PR.

Scoped hygiene run:

• git diff --cached --check
• ruff format --check tests/test_top_k_top_p_sampler.py
• ruff check tests/test_top_k_top_p_sampler.py
• clang-format --dry-run --Werror src/base/top_k_top_p_sampler.h src/native/ascend/ops/top_k_top_p_sampler/kernel.h
• Grep sweep for assert trailing periods, uppercase assert wording, and forbidden PR-comment phrase.

---

Checklist

Title, Branch, and Commits

• PR title follows Conventional Commits (e.g. feat(nvidia): …, fix(cuda/gemm): …).
• Branch name follows <type>/xxx-yyyy-zzzz where <type> matches the PR title's Conventional Commits type and words are joined with hyphens (see CONTRIBUTING.md §Branches).
• Each commit message follows Conventional Commits.
• Small PR is a single squashable commit; or, for a large PR, every commit is meaningful, well-formed, and independently reviewable (see CONTRIBUTING.md §Pull Requests).
• No stray merge commits from master — the branch is rebased cleanly on top of the current master.
• No fixup! / squash! / wip commits remain.

Scope and Design

• Changes are minimal — nothing unrelated to the stated motivation was added (CONTRIBUTING.md §Code/General).
• No dead code, commented-out blocks, debug prints, printf/std::cout/print(...) left behind, or TODO without an owner and issue link.
• No unrelated formatting churn that would obscure the diff.
• Public API changes (if any) are intentional, documented, and reflected in affected callers/tests.

General Code Hygiene (applies to all languages)

• The code is self-explanatory; comments were added only where the why is non-obvious (CONTRIBUTING.md §Code/General).
• Every modified or added file ends with a single trailing newline (CONTRIBUTING.md §Code/General).
• No trailing whitespace, tab/space mixing, or stray BOMs.
• Identifiers in comments and error messages are wrapped in backticks (e.g. the `seqlens_k` tensor) (CONTRIBUTING.md §Code/General).
• All comments and error messages are in English (CONTRIBUTING.md §Code/General).
• Comments and error messages are complete sentences — capitalized first letter, terminal punctuation — unless the language/framework convention says otherwise (CONTRIBUTING.md §Code/General; §Python).

C++ Specific (if C++ files changed)

• Code follows the Google C++ Style Guide strictly.
• clang-format (version 21, per .github/workflows/clang-format.yml) has been run against all modified .h, .cc, .cuh, and .mlu files; the diff is clean.
• clang-tidy concerns (per .clang-tidy) have been reviewed — no new warnings beyond the existing baseline.
• Operator parameter order is inputs first, outputs last; attributes are between inputs and outputs; naming follows PyTorch → ONNX → CUDA API precedence (CONTRIBUTING.md §C++).
• No exceptions are thrown. Error paths use assert with messages that include at least __FILE__, __LINE__, and __func__ (CONTRIBUTING.md §C++).
• Error and warning message wording follows the LLVM Coding Standards (CONTRIBUTING.md §C++).
• Kernel files are named correctly: custom = kernel / kernel_v2 / …; well-known algorithms use the algorithm name; library-based implementations use the library name (CONTRIBUTING.md §C++).
• N/A - Kernel and kernel launcher separation does not apply to this ACLNN wrapper header.
• Constructor initializer list order matches member declaration order (CONTRIBUTING.md §C++).
• Exactly one blank line between classes, between classes and functions, and between functions (CONTRIBUTING.md §C++).
• Exactly one blank line between members (functions and variables) within a class (CONTRIBUTING.md §C++).
• Exactly one blank line before and after the contents of a namespace (CONTRIBUTING.md §C++).
• New operators added via src/base/<op>.h (inheriting Operator<Op>) with platform implementations under src/<category>/<platform>/ inheriting the base (CONTRIBUTING.md §Adding an Operator).
• No raw new/delete; RAII / smart pointers / existing allocators are used.

Python Specific (if Python files changed)

• Code is PEP 8 compliant; ruff check passes cleanly on CI (see .github/workflows/ruff.yml).
• ruff format --check passes cleanly — if not, run ruff format and commit the result.
• N/A - No Python comments or docstrings were added.
• Framework-specific conventions (e.g. lowercase pytest.skip messages without terminal period) are honored where applicable (CONTRIBUTING.md §Python).
• No blank line between the function signature and the body when there is no docstring or comment (CONTRIBUTING.md §Python).
• A blank line is present before and after if, for, and similar control-flow statements (CONTRIBUTING.md §Python).
• A blank line appears before each return, except when it directly follows a control flow statement like if or for.
• N/A - No docstrings were added.
• N/A - Surrounding operator tests do not use type hints.

Testing

• pytest was run locally on every supported platform that this PR can affect, and the results are recorded in the "Test Results" table above (CONTRIBUTING.md §Pull Requests).
• For any platform that could not be tested, an explicit reason is given in the table and a reviewer with access has been tagged.
• New functionality has matching tests under tests/ following tests/test_add.py / tests/test_gemm.py patterns (CONTRIBUTING.md §Adding an Operator).
• Tests use pytest.mark.parametrize correctly: dependent parameters share one decorator (e.g. @pytest.mark.parametrize("dtype, rtol, atol", …)), independent parameters use separate decorators ordered by parameter declaration.
• Where appropriate, pytest.mark.auto_act_and_assert is used and the test returns a Payload whose func and ref share the same calling convention.
• Default dtype / device parameterization is relied on, or overridden with an explicit pytest.mark.parametrize when necessary.
• N/A - No flaky parallelism constraint was observed in the focused run.
• N/A - This is not a bug fix.

Build, CI, and Tooling

• The project builds cleanly from a fresh directory with pip install .[dev] on at least one affected platform.
• compile_commands.json still regenerates (CMake option CMAKE_EXPORT_COMPILE_COMMANDS=ON in pyproject.toml — required by the code-lint skill and clang-tidy -p).
• N/A - No new backend or device auto-detection was added.
• Only one CUDA-like GPU backend is selectable at a time — the existing mutual-exclusion check in CMakeLists.txt is not broken.
• Both CI workflows (clang-format.yml, ruff.yml) are green locally (or expected to be green on CI).
• No new runtime dependency was added without updating pyproject.toml's [project.optional-dependencies] (or justified in the PR description).

Documentation

• N/A - No README, CONTRIBUTING, build flag, or developer workflow behavior changed.
• New operators, new dispatch helpers, or new public utilities are documented (docstring, header comment, or an addition to CONTRIBUTING.md §Some Code Explanations).
• N/A - No user-visible breaking change.

Security and Safety

• No secrets, access tokens, internal URLs, customer data, or personal hardware identifiers have been committed.
• Third-party code is license-compatible and attributed.
• No unsafe pointer arithmetic, uninitialized reads, or missing bounds checks were introduced.