$1.$ Run the cuda$\_$hello program $($Program $6\mathrm{.}1)$ without the call to cudaDeviceSynchronize. What happens?

$2.$ When we ran cuda$\_$hello $($Program $6\mathrm{.}1)$ with $10$ threads, the output was ordered by the threads ranks: the output from thread $0$ was first, then the output from thread $1,$ then the output from thread $2,$ etc. Is this always the case? If not, what is the smallest number of threads you can start for which the output is not always ordered by the threads ranks? Can you explain why?

$3.$ When you run the cuda$\_$hello program $($Program $6\mathrm{.}1),$ what is the largest number of threads you can start? What happens if you exceed this number?