feat(cambricon): add ReLU provider

[baominghelly]

Summary

• Add a native Cambricon ReLU provider in src/native/cambricon/ops/relu/.
• Support contiguous and strided tensors, in-place execution, partial-overlap staging, integer dtypes, and PyTorch-compatible special-value semantics.
• Reuse the shared Cambricon task-range and logical-offset helpers from src/native/cambricon/kernel_utils.h.

Motivation

InfiniOps currently lacks a native Cambricon implementation for the canonical relu operator. This provider enables ReLU execution on Cambricon without relying on another backend.

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

Smoke Test Result

cmake -S . -B /workspace/build/pr-relu-20260901 \
  -DAUTO_DETECT_DEVICES=OFF \
  -DWITH_CPU=ON \
  -DWITH_CAMBRICON=ON \
  -DWITH_TORCH=OFF \
  -DWITH_LINKED=OFF \
  -DGENERATE_PYTHON_BINDINGS=ON \
  -DINFINI_OPS_OPS=relu \
  -DINFINI_RT_ROOT=/workspace/install/infinirt-master-test \
  -DPython_EXECUTABLE=/torch/venv3/pytorch/bin/python3.10 \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_INSTALL_PREFIX=/workspace/build/pr-relu-20260901-site/infini
cmake --build /workspace/build/pr-relu-20260901 --parallel "$(nproc)"
cmake --install /workspace/build/pr-relu-20260901

Result: targeted Cambricon build and install passed.

Test Results on Supported Platforms

Platform	Affected	Build / Smoke Result	Full Result / Notes
NVIDIA	No	N/A - not affected	N/A - not affected
Iluvatar	No	N/A - not affected	N/A - not affected
MetaX	No	N/A - not affected	N/A - not affected
Cambricon	Yes	Targeted build passed	python -m pytest -q tests/test_relu.py --devices cambricon --maxfail=1 -rs: 38 passed, 9 skipped
Moore	No	N/A - not affected	N/A - not affected
Ascend	No	N/A - not affected	N/A - not affected

Full `pytest` output (optional)

sssssss................................ss......                          [100%]
SKIPPED [7] tests/test_relu.py:55: Cambricon device code does not support float64 comparisons
SKIPPED [2] tests/test_relu.py:136: Cambricon device code does not support float64 comparisons
38 passed, 9 skipped in 0.31s

Benchmark / Performance Impact

N/A - no benchmark was run.

Notes for Reviewers

• Supported dtypes are float32, float16, bfloat16, int64, int32, int16, int8, and uint8.
• Float16 and bfloat16 values are converted to float32 in NRAM for comparison so NaN and signed-zero behavior matches PyTorch.
• Float64 is intentionally rejected because the Cambricon device compiler does not support the required float64 comparisons; the corresponding MLU cases are skipped explicitly.
• The implementation stages input data when input/output partially overlap and supports both contiguous and strided layouts.
• Repository-wide smoke tests were not run; this Draft PR records the targeted Cambricon build and complete tests/test_relu.py result.