This slide serves as a reminder that the actual hardware we will be using for the implementation is important. The previous algorithms for scan were optimized for either one core, or an infinite number of cores. However, in reality almost all systems we use will be neither of these. We thus have to consider the best implementation for the specific systems we will be using (in this case two cores). In this case, the added work for the infinite core algorithm just isn't worth it given only two cores.
This slide serves as a reminder that the actual hardware we will be using for the implementation is important. The previous algorithms for scan were optimized for either one core, or an infinite number of cores. However, in reality almost all systems we use will be neither of these. We thus have to consider the best implementation for the specific systems we will be using (in this case two cores). In this case, the added work for the infinite core algorithm just isn't worth it given only two cores.
This comment was marked helpful 0 times.