JRWynneIII / MatrixMagic

A scalable, CUDA accelerated linear algebra library
0 stars 1 forks source link

MatrixMagic

A scalable, CUDA accelerated set of linear algebra functions

This library is designed to make working with matricies and vectors simpler for the programmer. Simply #include the matrixMagic.h header file and use -L ./lib and -lMatrixMagic and they will provide you with access to the Matrix class and various other functions.

So far, this provides the user with:

And many more functions/features are in the works including a Conjugate Gradient routine to compute Ax = b.

As of now, all functions and objects are defined inside the main header file. This will be changed in the future to be more like a typical C++ library.

To recompile the MatrixMagic shared library, simply run:

$ make

To compile all the tests, run:

$ make tests

Files

matrix.txt -- Small test matrix
vector.txt -- Small test vector matrixMagic.h -- Contains the Matrix class
s3dkt3m2.dat -- Part of a matrix from the Matrix Market
makefile -- Allows you to make the shared library and to compile all the tests test/ -- Directory that contains test code for each of MatrixMagic's features and functions.