Computer Graphics Linear Algebra Library

Introduction

cglinalg is a low-dimensional linear algebra library aimed at real-time computer graphics, game engine development, and real-time collision detection. This library provides a strongly typed system for computer graphics applications.

The library is designed with the following goals in mind:

• Ergonomics -- The system should be easy to understand and use.
• Cross Platform -- It should be portable to other ecosystems like other C/C++ and Rust libraries. Every data type can be treated like a fixed-sized array so they can be sent to across FFI boundaries.
• Few Dependencies -- The library should be relatively self-contained. To support portability and maintainability, cglinalg is designed with few external dependencies. Also, the library only uses stable Rust.
• Type Safety -- Leverage Rust's type system and zero-cost abstractions to ensure code correctness, abstraction, and intelligibility do not come at the cost of performance.
• Flexibility -- The library should serve as a type-agnostic cornerstone for computer graphics applications. The data types in cglinalg are generic over their scalars so they can operate on multiple scalar types.
• Speed And Efficiency -- Operations should be fast and efficient. SIMD instructions and architecture specific optimizations should be used where possible. Moreover the library should compile fast.

Getting Started

To use the library in your project, add cglinalg as a dependency in your Cargo.toml file:

[dependencies]
cglinalg = "0.21.1"

After that, place the crate declaration in either your lib.rs file or your main.rs file

extern crate cglinalg;

The library aims to be as platform agnostic as possible. By default, the library supports any environment that supports the standard library std, but because the library does not require any allocations, it also supports environments built on either core or alloc. std is the default support feature, but you can add support for either alloc or core by adding

[dependencies.cglinalg]
# Use `cglinalg` with the `alloc` crate
features = ["alloc"]

for the alloc crate, or

features = ["core"]

for the core crate.

Features

cglinalg is a low-dimensional linear-algebra library aimed at specific application domains that make heavy use of computer graphics. It includes the most common linear algebra operations for implementing rendering algorithms, real-time collision detection, etc. All data types are designed to be exportable to external interfaces such as foreign function interfaces or external hardware. This serves the cglinalg goal to be a platform agnostic foundation for computer graphics applications in other languages and ecosystems as well. Specific features of the library include:

• Basic linear algebra with matrices and vectors up to dimension four.
• Quaternions, Euler angles, and rotation matrices for doing rotations.
• All data types are parametrized to work over a large range of numerical types.
• An optional transformation system for working with affine and projective transformations on points and vectors. This library distinguishes points from vectors and locations in space vs. displacements in space. This matters when working with working with affine transformations in homogeneous coordinates.
• Transformations including translation, reflection, shear, scale, and rotation operations.
• Orthographic projections and perspective projections for camera models.
• Typed angles and typed angle trigonometry that statically guarantee that trigonometry is done in the right units.
• The library makes heavy use of property testing ensure code correctness.
• The library makes heavy use of statistical microbenchmarking.