#complex-numbers #matrix-vector #matrix #vector #matrix-operations

accelerate-general

A general matrix and vector operations library using Apple's Accelerate framework

1 unstable release

0.1.0 Oct 12, 2024

#1156 in Math

22 downloads per month

MIT/Apache

255KB
1.5K SLoC

Accelerate-General: Matrix and Vector General Operations using Apple's Accelerate Framework in Rust

This project provides Rust FFI bindings to interact with Apple's Accelerate framework, enabling efficient matrix and vector operations with both single-precision (f32) and double-precision (f64) floating-point numbers, as well as complex numbers (Complex<f32>, Complex<f64>).

Features

  • Vector and Matrix Operations: Provides efficient routines for basic linear algebra operations.
  • Single and Double Precision: Support for both single (f32) and double (f64) precision operations.
  • Complex Number Support: Complex arithmetic with both single-precision and double-precision complex numbers.
  • Optimized for Apple Platforms: Uses the Accelerate framework for high performance on macOS and iOS.

Requirements

  • Rust (1.60 or higher)
  • Apple's Accelerate Framework (iOS 16.4+ ,iPadOS 16.4+ ,Mac Catalyst 16.4+ ,macOS 13.3+ ,tvOS 16.4+ ,visionOS 1.0+ ,watchOS 9.4+)
  • FFI for interfacing with C functions

Dependencies

The project uses the following dependencies:

  • num-complex: To support complex number operations in Rust.
  • std::ffi: For calling C functions via FFI.
[dependencies]
num-complex = "0.4"

Getting Started

Installation

Clone the repository and include it in your project by adding the following to your Cargo.toml:

[dependencies]
accelerate-general = { path = "/path/to/your/cloned/repo" }

Usage

  1. Import the required modules and types from the library.
  2. Use FFI functions for matrix and vector operations.

Safety

All functions in this library are marked as unsafe since they directly interface with C libraries via FFI. It is the responsibility of the caller to ensure that:

  • Memory pointers passed to the functions are valid and aligned.
  • vectors and matrices are correctly dimensioned and non-null.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

This project uses Apple’s Accelerate framework for optimized matrix and vector operations. The Accelerate framework provides high-performance BLAS routines that are used via FFI in this project.

Dependencies

~270KB