EECS-338-Final-Project Power iteration for large dense matrices using multiple GPUs The idea here is the following: Split up matrix vector product row wise Compute each part of the larger matrix-vector product on GPUs Observe results (i.e. convergence etc.) on master