Statslabs::Matrix - The C++ Linear Algebra Library (in development)

The repository contains the Statslabs::Matrix C++ linear algebra library code. Statslabs::Matrix is the fundamental package of Statslabs for statistical computing in C++.

Overview

Statslabs::Matrix is the fundamental package of Statslabs for statistical computing in C++. The Statslabs.Matrix library code is based on the matrix design chapter in 'The C++ Programming Language (4th Edition)' and provides:

• A Matrix Template: Construction and Assignment; Subscripting and Slicing
• Matrix arithmetic operations: Scalar Operations; Additions; Multiplication
• Matrix Implementation: slice; MatrixSlice; MatrixRef; Matrix List Initialization; Matrix Access; Zero-Dimensional Matrix
• An interface to Intel(R) MKL BLAS operations which apply to the Matrix template

Prerequisites

CMake >= 3.6.0
Intel Math Kernel Library (Intel MKL)

Installation on Ubuntu / macOS

1. Clone the repository.

   git clone git@github.com:statslabs/matrix.git

2. Configure the project.

   cd matrix
   mkdir build && cd build
   cmake ..

   To build docs/examples/tests, configure matrix with the following cmake options:

   -DSLAB_INCLUDE_DOCS=ON
   -DSLAB_INCLUDE_EXAMPLES=ON
   -DSLAB_INCLUDE_TESTS=ON

3. Compile and install the library.

   make
   make install

Example program

examples/demo.cc:

#include <iostream>
#include "slab/matrix.h"

using namespace std;
using namespace slab;

int main() {
  mat A = {
      {1, 2, 3},
      {4, 5, 6},
      {7, 8, 9}
  };

  mat B = {
      {1, 2, 3},
      {4, 5, 6},
      {7, 8, 9}
  };

  // Element-wise addition
  cout << "A + B = " << A + B << endl;

  // Element-wise subtraction
  cout << "A - B = " << A - B << endl;

  // Element-wise multiplication
  cout << "A * B = " << A * B << endl;

  // Element-wise division
  cout << "A / B = " << A / B << endl;

  // Matrix multiplication
  cout << "matmul(A, B) = " << matmul(A, B) << endl;

  return 0;
}

Integration of Matrix in your own project

To make the project simple enough, we will create a CMake project for demo.cc.

1. Make a project folder.

   mkdir example && cd example

2. Create demo.cc and CMakeLists.txt in the project folder where file CMakeLists.txt should look like:

   cmake_minimum_required(VERSION 3.0)
   project(example)
   set(CMAKE_CXX_STANDARD 11)
   add_executable(example demo.cc)
   
   find_package(Matrix REQUIRED)
   target_link_libraries(example Statslabs::Matrix)

3. Perform a out-of-source build.

   mkdir build && cd build
   cmake ..
   make

4. Run the program.

   ./example