This directory contains code for benchmarking different matrix-vector
product functions using OpenMP parallelism.

It supports both dense and sparse matrices.

Files:
        matrix.{h,cpp}  Representation and manipulation of matrices
        mbench.{h,cpp}  Benchmarking code
        mvmul.cpp       Various implementations of matrix-vector products
        CycleTimer.h    Timing code
        main.cpp        Main function

        results.txt:    Results from running on 8-core CPU

Running 'make' generates a program mrun.

Running './mrun -h' displays runtime options.