Sparse Matrix-Vector Multiplication using Pthreads

Optimizations: 1) Loop Optimization [1]: A typical matrix-vector multiplication (matrix in CSR format) consists of a nested loop where the outer loop iterates over all the rows and the inner loop iterates over columns in those rows. Since the data is stored in a sequential fashion in CSR (one row after the other), the data can be accessed by the nested loop using a single loop variable instead of two (in the typical algorithm). As expected, loop unrolling gave a good performance boost to our initial simple CSR-based algorithm.