Build, C++ And Platforms
Use this page when you need native builds, C++ embedding, or platform-specific binaries.
Most users start with:
pip install blazerulesUse this page when you need to embed the C++ library, tune a source build, produce your own container image, or target a specific CPU family.
Source Build
cmake --preset linux-x86_64-release-dispatch
cmake --build --preset linux-x86_64-release-dispatch -jOther presets:
| Preset | Target |
|---|---|
macos-arm64-release | Apple Silicon |
linux-arm64-release-neon | Linux ARM64 / Graviton |
linux-x86_64-release-dispatch | Linux x86_64 runtime dispatch |
linux-x86_64-release-avx512 | Linux x86_64 with AVX-512 kernels |
cloud-portable-release | Mixed cloud CPU fleet |
windows-x64-release-dispatch | Windows x64 |
C++ Embedding
#include <blazerules/engine.h>
blazerules::RuleEngine engine;
engine.load_rules("rules.yaml");
auto result = engine.evaluate_ndjson(payload);See C++ API and Embed In C++.
Python Extension
The Python module is pybind11 over the same C++ engine:
import blazerules
engine = blazerules.RuleEngine()See Python API.
Platform Notes
- ARM64 macOS / Linux: NEON backend.
- x86_64 Linux / Windows: AVX2 runtime-dispatched backend where available.
- AVX-512 kernels are compiled in full builds and selected only when supported.
- For mixed cloud fleets, prefer runtime-dispatch presets over a binary tuned for one machine.