Explore Performance 2

[NullVoxPopuli]

Related:
#1596

The goal of this PR is to try to see if we can undo the perf losses reported by

• Rendering performance regression between 0.84.3 and 0.85.1? #1590

[github-actions]

duration phase estimated improvement -110ms [-168ms to -48ms] OR -0.76% [-1.17% to -0.33%]
renderEnd phase no difference [-2ms to 1ms]
render1000Items1End phase estimated improvement -7ms [-14ms to 0ms] OR -0.58% [-1.28% to 0%]
clearItems1End phase no difference [-2ms to 5ms]
render1000Items2End phase no difference [-15ms to 0ms]
clearItems2End phase no difference [0ms to 5ms]
render5000Items1End phase estimated improvement -33ms [-59ms to -5ms] OR -0.82% [-1.45% to -0.12%]
clearManyItems1End phase no difference [-1ms to 1ms]
render5000Items2End phase estimated improvement -34ms [-66ms to -1ms] OR -0.91% [-1.78% to -0.02%]
clearManyItems2End phase no difference [-2ms to 1ms]
render1000Items3End phase no difference [-6ms to 16ms]
append1000Items1End phase no difference [-12ms to 12ms]
append1000Items2End phase no difference [-36ms to 1ms]
updateEvery10thItem1End phase no difference [-11ms to 2ms]
updateEvery10thItem2End phase no difference [0ms to 4ms]
selectFirstRow1End phase no difference [0ms to 0ms]
selectSecondRow1End phase no difference [0ms to 0ms]
removeFirstRow1End phase no difference [0ms to 0ms]
removeSecondRow1End phase no difference [0ms to 0ms]
swapRows1End phase no difference [0ms to 0ms]
swapRows2End phase no difference [0ms to 0ms]
clearItems4End phase no difference [-7ms to 0ms]
paint phase no difference [0ms to 0ms]

[23:27:41] Generating Benchmark Reports [started]
[23:27:46] Generating Benchmark Reports [completed]

Benchmark Reports    

JSON: /home/runner/work/glimmer-vm/glimmer-vm/tracerbench-results/compare.json

PDF: /home/runner/work/glimmer-vm/glimmer-vm/tracerbench-results/artifact-1.pdf

HTML: /home/runner/work/glimmer-vm/glimmer-vm/tracerbench-results/artifact-1.html

[NullVoxPopuli]

duration phase estimated improvement -110ms

woah, that's nice -- Suppose I should compare with the other versions

[NullVoxPopuli]

The improvement revealed by tracerbench translates all they way through to ember.

[NullVoxPopuli]

Closing, because #1606 solved a lot