RaftLib/RaftLib

Health v2 · maintenance only

Adoption not part of the score

999 stars · 125 forks observed · 2026-10-21

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-07, confidence 0.95

RaftLib is an open-source C++ library for building parallel and concurrent data processing pipelines using a stream/dataflow programming model. Compute kernels are linked together with iostream-like shift operators to form DAGs, with lock-free FIFO channels and automatic scheduling handling parallelism without manual threading.

Use cases

• build high-performance data processing pipelines in C++
• parallelize streaming computations without pthreads or OpenMP
• construct directed acyclic graphs of compute kernels
• implement concurrent dataflow applications for HPC
• process data streams with automatic load balancing and backpressure
• replace hand-written thread management with deterministic pipelines

When to choose

• you need parallel dataflow pipelines in C++ with a simple operator-based API
• you want to avoid explicit threading primitives like std::thread or OpenMP
• you're building high-performance computing or stream processing applications
• you need lock-free inter-kernel communication with auto-parallelization

When to avoid

• you need distributed multi-node processing rather than single-machine parallelism
• your project is not C++
• you need a full streaming platform like Kafka or Flink with clustering and persistence
• you require GPU-specific dataflow frameworks

Facets

library · maturity active

concurrency streaming middleware developer-tools performance microservices big-data cpp cross-platform dataflow-programming parallel-computing hpc pipelines kernels lock-free c-plus-plus linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="RaftLib/RaftLib")