1. Consider that you are writing (a) an application that performs independent calculations that can be executed concurrently (e.g., matrix multiplication and sorting of different data) and (b) an application that performs calculations and I/O operations that can be executed concurrently (e.g., sorting data from a file and saving the sorted date into another file). What would be more efficient to implement: user-level threads or kernel-level threads? Discuss each answer for a uniprocessor machine and a multiprocessor machine.