Efficient Tensor Operations Via Compression And Parallel Computation