This PR introduces a new **multi-core pipeline controller** that manages and monitors the execution of pipelines within the current process. The controller uses a **thread-per-core model, with each thread pinned to a specific CPU core** to maximize multi-core efficiency and reduce contention, ensuring predictable and optimal CPU usage. ## Other improvements * Added a new CLI command to run pipelines for demo purposes * Switched to MiMalloc for better memory efficiency and reduced fragmentation compared to the system allocator * Made `PipelineConfig` `Send` to allow sharing across cores; other data remains non-`Send` * Introduced a new `Quota` section in the `PipelineGroup` configuration, currently supporting a maximum core count for the engine (foundation for future resource control features) * Enhanced the Control Message System: * Pipeline control messages: New system for nodes to send control messages to the pipeline * Renamed previous system: Former control messages are now "node control messages" for clarity * Implemented TimerTick control message for timer-based control messaging * Created demo-specific pipeline configurations ## Future enhancements (upcoming PRs) * #872 * #871 * #831 * #833 ## Preliminary benchmark results <img width="720" height="541" alt="image" src="https://github.com/user-attachments/assets/a80ebfe8-01c5-4488-96bd-d376bfb887cb" />

HTTP endpoints to expose metrics, status and health checks #872

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions