perf: add js-framework-benchmark tests

[nolanlawson]

Details

As we track our progress with fine-grained reactivity (#3624), I think it would be helpful to have the actual js-framework-benchmark benchmarks in our local benchmarks. Otherwise, it's kind of a pain to set up js-framework-benchmark itself, link it to LWC, and run it to see the results.

This PR is my attempt to replicate as closely as possible what js-framework-benchmark is doing. A few limitations:

• There is some overlap here with existing benchmarks, which were clearly also inspired by js-framework-benchmark but are slightly different.
• Some of the benchmarks execute quite quickly (<5ms), which is not great for Tachometer since it tends to lead to lots of variance. (Update: opened perf: add throttling to tachometer benchmarks #4070 to resolve this.)

Despite these limitations, I still think this is worth doing, at least to see our historical trend over time as we try to improve these benchmark results.

Does this pull request introduce a breaking change?

• 😮‍💨 No, it does not introduce a breaking change.

Does this pull request introduce an observable change?

• 🤞 No, it does not introduce an observable change.

[nolanlawson]

All of these tests are adapted from https://github.com/krausest/js-framework-benchmark/blob/6c9f43f62604faa42bc7ddce634fde5a2e8b6eb7/webdriver-ts/src/benchmarksPuppeteer.ts

[nolanlawson]

We could extract out these querySelectors so that we don't pay the cost in the actual benchmark run, but 1) it doesn't seem to affect the numbers much, and 2) js-framework-benchmark seems to be implicitly doing this as well (assuming that WebDriver/Puppeteer delegate out to document.querySelector).

[jmsjtu]

LGTM, thanks for adding these!