Parallel Sparse Matrix Computations Using the PINEAPL Library: A Performance Study