The PriorityTicketHolder will not have comparable performance to the SemaphoreTicketHolder when the priority of operations are varied. This is because low priority operations will always wait for the normal priority queue to be cleared in the contended case. 

Since microbenchmarks are relatively cheap, one option is to have two variations of the PriorityTicketHolder benchruns:
(1) Where all operations are normal - this allows for a more reasonable direct comparison to SemaphoreTicketHolder performance since the SemaphoreTicketHolder disregards ticket priority
(2) Where operations are varied low and high - this will allow for iterative comparisons of the PriorityTicketHolder's performance when it manages a balanced load of low and high priority operations