tools: add benchmark runner to test-shared

[aduh95]

• it allows non-collaborators can run them on their forks.
• testing on different arch gives more relevant results.
• if folks want to run benchmarks on their own machine, I find it helpful to have a GHA workflow to copy to understand how to set it up.

Example of benchmark workflow: https://github.com/aduh95/node/actions/runs/18478104954

[nodejs-github-bot]

Review requested:

• @nodejs/actions

[targos]

I would not be very confident running benchmarks on Cloud VMs that we don't know much about. That's why we have dedicated bare metal machines for that in CI.

[aduh95]

I would not be very confident running benchmarks on Cloud VMs that we don't know much about. That's why we have dedicated bare metal machines for that in CI.

Do you think we should add a comment about that in the output? This probably goes without saying, but I'm not suggesting we get rid of the bare metal machines, I'm suggesting that as an addition.

[targos]

Maybe. Something that says only the results of the Jenkins benchmark CI can be used make a decision.

[aduh95]

One upside of having something like that: it allows one to validate changes to the Benchmark tooling during a CI lockdown