Crate metal_matrix 

Metal Matrix

A high-performance linear algebra library with Metal GPU acceleration.

This library provides GPU-accelerated matrix operations using Apple’s Metal framework. It’s designed for efficient computation of common linear algebra operations like matrix multiplication, addition, subtraction, transposition, and scalar multiplication.

Features

• GPU-accelerated matrix operations
• Clean, ergonomic API
• Support for vectors as 1D matrices
• CPU fallback implementations
• Comprehensive error handling

Example

use metal_matrix::{MetalContext, Matrix, matrix_multiply};
use anyhow::Result;

fn main() -> Result<()> {
    // Initialize Metal context
    let context = MetalContext::new()?;
     
    // Create matrices
    let mut a = Matrix::new(3, 2);
    let mut b = Matrix::new(2, 4);
     
    // Fill matrices with data
    // ...
     
    // Multiply matrices
    let result = matrix_multiply(&context, &a, &b)?;
     
    Ok(())
}

Re-exports

pub use matrix::Matrix;

pub use metal_context::MetalContext;

pub use operations::*;

Modules

kernels

Metal kernel definitions and paths

matrix

Matrix data structure and methods

metal_context

Metal context and device management

operations

Matrix operations implementation