A principle stating that as the number of processors increases, the size of the problem can be scaled proportionally, allowing greater overall speedup than predicted by Amdahl's Law because the parallel portion of the workload grows with problem size.
