bphun / MatrixLibrary

Hardware accelerated C++11 Matrix Library

Matrix Library

This is a basic C++11 matrix library that supports the operations listed below:

• Scalar addition
• Matrix addition
• Scalar subtraction
• Matrix subtraction
• Scalar multiplication
• Matrix multiplication
• Scalar division

Experimental Features:

• OpenCL GPU acceleration
• OpenMP support

Current features in development:

• Cuda GPU acceleration
• Determinant calculation
• Matrix inverse computation
• Matrix division

Building the library

To compile the matrix library in single threaded computation mode, follow these steps:

git clone https://github.com/bphun/MatrixLibrary.git
cd MatrixLibrary
mkdir build
cd build
cmake ../
make

To compile the matrix library in multi-threaded computation mode(using OpenMP), ensure that you have a compiler installed that supports OpenMP then follow these steps:

git clone https://github.com/bphun/MatrixLibrary.git
cd MatrixLibrary
mkdir build
cd build
cmake -DCMAKE_CXX_COMPILER=<Path to compiler with OpenMP support> -DADDITIONS=OPENMP ../
make

To compile the matrix library in GPU computation mode(using OpenCL), ensure that you have a compiler installed that supports OpenMP then follow these steps:

git clone https://github.com/bphun/MatrixLibrary.git
cd MatrixLibrary
mkdir build
cd build
cmake -DADDITIONS=OPENCL ../
make

Usage

Once this matrix library is linked with your project, you can include the library in your source with #include <Matrix.h>.

Examples

Although all of the following examples use matrices that hold integers, this matrix library can be used with all numeric data types.

Instantiating an empty matrix:

Matrix<int> matrix;

Instantiating an empty matrix with a certain number of rows and columns:

const int rows = 1024;
const int columns = 1024;
Matrix<int> matrix(rows, columns)

Instantiating a matrix intialized with values of a numeric array:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};
Matrix<int> matrix(intArray, rows, cols);

Instantiating a matrix intialized with values of a numeric vector:

const int rows = 2;
const int cols = 2;

vector<int> vector{1, 2, 3, 4};

Matrix<int> matrix(vector, rows, cols);

Adding two matrices:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};

Matrix<int> matrixA(intArray, rows, cols);
Matrix<int> matrixB(intArray, rows, cols);

Matrix<int> &resultMatrix = matrixA + matrixB;

Multiplying a matrix by a scalar:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};

Matrix<int> matrix(intArray, rows, cols);

Matrix<int> &resultMatrix = matrix * 2;

Multiplying two matrices:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};

Matrix<int> matrixA(intArray, rows, cols);
Matrix<int> matrixB(intArray, rows, cols);

Matrix<int> &resultMatrix = matrixA * matrixB;

Subtracting two matrices:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};

Matrix<int> matrixA(intArray, rows, cols);
Matrix<int> matrixB(intArray, rows, cols);

Matrix<int> &resultMatrix = matrixA - matrixB;

Dividing a matrix by a scalar:

const int rows = 2;
const int cols = 2;

int intArray[rows * cols] = {1, 2, 3, 4};

Matrix<int> matrix(intArray, rows, cols);

Matrix<int> &resultMatrix = matrix / 2;