Make changes suggested by @sokra to optimize chunk graph generation

[MLoughry]

Addresses concerns in #8126 by attempting to implement suggestions from @sokra to optimize Compilation.processDependenciesBlocksForChunkGroups.

1. Filter the dependencies without modules within the for loop, rather than using .filter() to allocate a new array (lines 1870-1879)
2. Re-use availableModule sets that only have one entry, rather than copying them (lines 1844 and 1900)
3. Avoid expensive areModulesAvailable() call when a chunk has already survived a filter. (lines 1818-1823).

I've probably mangled sokra's ideas beyond recognition, but opening this PR so he can more easily comment on what I should do. Currently, I'm not actually seeing perf improvements from these changes.

What kind of change does this PR introduce?

Performance improvement (without any intended behavioral change)

Did you add tests for your changes?

No; as a perf change with no behavior change, I rely on the existing tests

Does this PR introduce a breaking change?

No

What needs to be documented once your changes are merged?

N/A

[webpack-bot]

For maintainers only:

• This need to be documented (issue in webpack/webpack.js.org will be filed when merged)

[webpack-bot]

Thank you for your pull request! The most important CI builds succeeded, we’ll review the pull request soon.

[sokra]

Thanks