Crate magetypes 

magetypes

Token-gated SIMD types with natural operators.

This crate provides SIMD vector types (f32x8, i32x4, etc.) that use archmage tokens for safe construction.

Supported Platforms

• x86-64: x86-64-v3 (128-bit, 256-bit), AVX-512 (512-bit)
• AArch64: NEON (128-bit)
• WASM: SIMD128 (128-bit) - compile with RUSTFLAGS="-C target-feature=+simd128"
• All other targets: Scalar fallback (pure array math, no hardware SIMD)

Example

use archmage::{X64V3Token, SimdToken};
use magetypes::simd::f32x8;

if let Some(token) = X64V3Token::summon() {
    let a = f32x8::splat(token, 1.0);
    let b = f32x8::splat(token, 2.0);
    let c = a + b;  // Natural operators!
    println!("Result: {:?}", c.to_array());
}

Re-exports

pub use archmage;

Modules

cast

Cross-tier casting utilities.

prelude

Platform-appropriate SIMD types without manual #[cfg] blocks.

simd

Token-gated SIMD types with natural operators.

Traits

SimdTypes

Trait that associates SIMD vector types with a capability token.

WidthDispatch

Trait providing access to all SIMD sizes from a capability token.