Can we improve on the performance of pop()?

[davecom]

I came across this article in which the performance of SwiftPriorityQueue was far lower than both an Objective-C priority queue and the C++ STL priority queue because we're spending 68% of his tests in pop() whereas they are spending <10% of their time in his tests on pop(). What can be improved about our pop() implementation? Here's the article, and the relevant section is 6.2:
https://miun.diva-portal.org/smash/get/diva2:1135549/FULLTEXT01.pdf

[davecom]

Closed by #20