nmatrix

NMatrix, also styled as nmatrix, is a matrix and multi-dimensional array library for the Ruby programming language. It is a key component of the SciRuby project, created to provide high-performance numerical data structures and operations for scientific computing in Ruby. The library offers dense and sparse representations and aims to combine Ruby’s usability with efficient, lower-level implementations.

NMatrix provides multi-dimensional arrays with customizable data types, including common numeric types such as floating-point and integer scalars, as well as boolean and complex numbers. It supports a range of shapes and supports indexing, slicing, and various element-wise and aggregate operations. In addition to basic arithmetic, the library exposes a collection of linear algebra and numerical methods, depending on the active backend and available dependencies. Sparse matrices are supported to help with memory efficiency in problems with large, mostly empty matrices.

Implementation and backends are centered on a native Ruby C extension, which enables performance improvements over pure Ruby code. The backend abstraction allows NMatrix to leverage optimized numerical libraries when they are available, and to interoperate with other components in the Ruby scientific stack. The design emphasizes compatibility with Ruby’s object model and with the broader SciRuby ecosystem.

In practice, NMatrix is used as a core building block for numerical computing in Ruby, providing a familiar array-like API while supporting more advanced linear algebra and scientific workflows. It is commonly paired with other SciRuby gems for data analysis, visualization, and statistical computing. See also SciRuby and NArray for related Ruby numerical libraries.