perf(run): improve yarn run performance

[paul-soporan]

What's the problem this PR addresses?

Helps a bit with #2575.

Testing it on https://github.com/jest-community/eslint-plugin-jest (CC @SimenB), it seems to save ~345.7 ms.

This PR doesn't change anything linker-specific, so it should also slightly optimize yarn run with the PnP linker.

Because #2576 overlaps with this PR in some places, the combined perf improvements of the 2 PRs will probably only save something like ~350-400ms.

How did you fix it?

• Parallelize the makePathWrappers.
• Bypass the linkers (and the ZipOpenFS) if the package is a workspace.
• Don't compute the packageAccessibleBinaries inside executePackageAccessibleBinaries again, as they are already computed inside the execute method of the yarn run command.
• Parallelize the async part of the loop in getPackageAccessibleBinaries - the remaining part still has to be sequential to guarantee consistent overwrite order of the binaries.

Checklist

• I have read the Contributing Guide.

• I have set the packages that need to be released for my changes to be effective.

• I will check that all automated PR checks pass before the PR gets reviewed.