My machine has 4/8 cores

From results, improvement from 1->2 threads and 2->4 threads is almost 2 times,
but not exactly, since creating threads and joining them takes time as well.

When the number of threads is larger then the number of physical cores of my CPU,
e.g. 4->10 threads, 10->100 threads, the improvement is insignificant, since two
threads cannot run on the same core at the same time.