ArchX — Your CPU already knows how to be fast. Let it decide.

ArchX is a mission-critical, adaptive acceleration engine for the modern Rust ecosystem. It eliminates the guesswork from performance optimization by dynamically routing workloads across SIMD, Multithreading, and GPU based on real-time hardware topology.

⚡ Why ArchX?

In the current landscape, software either settles for "good enough" performance or drowns in unsafe SIMD intrinsics and manual thread tuning. ArchX changes the narrative.

The Problem

Traditional libraries (like Rayon) are excellent for data parallelism, but they don't care about your hardware's specific SIMD breadth (AVX2 vs AVX-512) or PCIe transfer overhead compared to CPU throughput.

The ArchX Solution

Adaptive Heuristics: Decision-making based on L1/L2 cache sizes, memory bus throughput, and core topology.
Zero-Dependency Core: Only std and serde (optional). No heavy runtime bloat.
Fluent Ecosystem: A professional DX that lets you optimize in seconds, not hours.

🚀 Quick Start (30 Seconds)

Basic Usage

The "Just Work" API for standard workloads.

use archx::add;

fn main() {
    let a = vec![1.0; 1_000_000];
    let b = vec![2.0; 1_000_000];
    let mut out = vec![0.0; 1_000_000];

    // ArchX detects your CPU and decides the best path (SIMD or Parallel).
    add(&a, &b, &mut out);
}

Advanced Fluent API (v1.2.0)

Professional control for enterprise-grade applications.

use archx::engine;

fn main() {
    engine()
        .with_profile(true)        // Enable high-fidelity diagnostics
        .with_limits(0.4)          // Stay under 40% CPU usage
        .with_auto_gpu()           // Offload to GPU if PCIe overhead is justified
        .execute(|| {
            // Your heavy compute logic
        });
}

🏗️ Ecosystem Integration

Integration	Guide
🚀 Tauri	Building high-performance desktop apps with ArchX
🌐 WASM	SIMD-accelerated browser compute
⏳ Async	Non-blocking parallel processing for servers

📊 Comparison: ArchX vs The World

Feature	Raw SIMD	Rayon	ArchX
DX	Hard (Unsafe)	Easy	Fluent
Adaptive	No	Partially	Yes (Smart)
Hardware Aware	Manual	No	Yes (Automatic)
GPU Sync	No	No	Yes

🛡️ Trust & Reliability

Safe by Design: 100% abstract over unsafe SIMD intrinsics.
Deterministic: Heuristics are calibrated for stability, not just peak bursts.
Production Proven: Benchmarked extensively on various hardware sets.

📅 v1.2.0 Roadmap

Plugin System: custom optimizer plugins for specialized domains.
Dynamic Re-calibration: live heuristic tuning based on system load.
ARM Neon parity: full-speed optimizations for Apple Silicon.

🤝 Contribution

Designed with love by Codevora Studio. We welcome contributors! Check CONTRIBUTING.md to get started.

archx 1.2.0