ringkernel

Main facade crate for the RingKernel GPU-native persistent actor model framework.

Overview

This crate re-exports the entire RingKernel API, providing a single entry point for users. It combines core abstractions, backend implementations, and derive macros into a unified interface.

Installation

[dependencies]
ringkernel = "0.1"
tokio = { version = "1", features = ["full"] }

For GPU backends:

# NVIDIA CUDA
ringkernel = { version = "0.1", features = ["cuda"] }

# WebGPU (cross-platform)
ringkernel = { version = "0.1", features = ["wgpu"] }

# All backends
ringkernel = { version = "0.1", features = ["all-backends"] }

Quick Start

use ringkernel::prelude::*;

#[tokio::main]
async fn main() -> std::result::Result<(), Box<dyn std::error::Error>> {
    // Create runtime with CPU backend
    let runtime = RingKernel::builder()
        .backend(Backend::Cpu)
        .build()
        .await?;

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

    // Kernel lifecycle
    kernel.deactivate().await?;
    kernel.activate().await?;
    kernel.terminate().await?;

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

Features

Re-exported Crates

• ringkernel-core - Core traits and types
• ringkernel-cpu - CPU backend
• ringkernel-cuda - CUDA backend (optional)
• ringkernel-wgpu - WebGPU backend (optional)
• ringkernel-derive - Proc macros

Examples

The crate includes 20+ examples:

# Basic usage
cargo run -p ringkernel --example basic_hello_kernel

# Messaging patterns
cargo run -p ringkernel --example kernel_to_kernel
cargo run -p ringkernel --example pub_sub

# CUDA code generation
cargo run -p ringkernel --example global_kernel
cargo run -p ringkernel --example stencil_kernel

# WebGPU
cargo run -p ringkernel --example wgpu_hello --features wgpu

License

Apache-2.0