perf_hooks: reducing overhead of performance observer entry list

[H4ad]

Continuing the work started on nodejs/performance#109.

                                                       confidence improvement accuracy (*)   (**)  (***)
perf_hooks/performance-observer.js pending=1 n=100000         ***    147.63 %       ±3.89% ±5.23% ±6.89%
perf_hooks/performance-observer.js pending=10 n=100000        ***    104.54 %       ±5.69% ±7.60% ±9.92%

Be aware that when doing many comparisons the risk of a false-positive
result increases. In this case, there are 2 comparisons, you can thus
expect the following amount of false-positive results:
  0.10 false positives, when considering a   5% risk acceptance (*, **, ***),
  0.02 false positives, when considering a   1% risk acceptance (**, ***),
  0.00 false positives, when considering a 0.1% risk acceptance (***)

cc @nodejs/performance

[mscdex]

I don't know if it's a common case or not, but it may/may not be worth guarding the sort with a conditional that checks if entries.length > 1 before sorting.

[H4ad]

I added, I got 8% faster for 1 item but degraded the else path by ~8%, so I didn't think that was worth it since probably this queue will be called with more items than 1.

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/54464/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/54514/

[nodejs-github-bot]

Landed in 328bdac