Welcome To MatX’s Documentation!#
MatX is a modern C++ library for numerical computing on NVIDIA GPUs and CPUs. The main features include:
Compile-time expression evaluation for generating GPU kernels
Near-native performance for GPU kernels while using a syntax similar to Python or MATLAB
Easy frontend API to many popular CUDA libraries
Easy frontend API to CPU libraries (NVPL, FFTW, OpenBLAS, BLIS) on ARM and x86
Header-only without need for compilation
Single tensor type used across entire API
Intuitive error messages
To start building tests, examples, and benchmarks, visit the Build Integration guide. This page also lists the requirements for each type of build.
If you’re new to MatX and want to jump right in, we recommend starting with the Quick Start guide. This guide walks through the MatX concepts and primary API functions to give you the tools to make your first application using MatX.
If you prefer learning by looking directly at complete examples, take a look in the examples
directory. Some
examples have full walk-throughs in the Example Applications section of the documentation.
Lastly, an API Reference guide is available for every function in MatX.
Note
We strive to make sure that every function in the MatX API has at least one example and one unit test. All example code in the API section is cross-referenced from the unit tests. This means that the examples in this documentation should always be correct and unit tests would catch any errors before they’re introduced or modified in the docs.
License#
MatX is released under the BSD-3-Clause license. Several external packages are used optionally and may have different license requirements. Please see the Build Integration guide for a list of external requirements.
Table of Contents#
- Quick Start
- Build Integration
- Build Options
- MatX Library Linking
- MatX in Offline Environments
- Basics
- API Reference
- Creating Operators and Tensors
- Manipulation Functions
- Input/Output
- Mathematical Functions
- Linear Algebra
- Statistics
- Searching and Sorting
- Logic Functions
- Casting Operators
- Windowing
- Signal and Image Processing
- Synchronization
- Polynomials
- Random Number Generation
- Discrete Fourier Transform
- Visualization
- Type Traits
- Example Applications
- Executor Compatibility
- Version Info