RingKernel

A GPU-native persistent actor model framework for Rust. Build high-performance GPU-accelerated actor systems with persistent kernels, lock-free message passing, and hybrid logical clocks.

API Documentation Getting Started View on GitHub

🚀 Persistent GPU Kernels

Long-running GPU kernels that maintain state across invocations, eliminating launch overhead.

🔗 Lock-free Message Passing

High-performance ring buffers for host-GPU and kernel-to-kernel communication.

⏱ Hybrid Logical Clocks

Causal ordering across distributed operations with HLC timestamps.

📦 Zero-Copy Serialization

Efficient message passing with rkyv-based zero-copy serialization.

💻 Multi-Backend Support

Target CPU, CUDA, WebGPU, or Metal with a unified API.

🎯 Proc Macro DSL

Declarative message and kernel definitions with derive macros.

use ringkernel::prelude::*;

async fn main() -> std::result::Result<(), Box<dyn std::error::Error>> {
    // Create runtime with auto-detected backend
    let runtime = RingKernel::builder()
        .backend(Backend::Auto)
        .build()
        .await?;

    // Launch a persistent kernel
    let kernel = runtime.launch("processor", LaunchOptions::default()).await?;

    // Kernel is active and ready
    println!("State: {:?}", kernel.state());

    runtime.shutdown().await?;
    Ok(())
}
            