Increase the benchmark warmup time to avoid warmup artifact

[JacobEvelyn]

This commit increases the warmup time used in our benchmark so that two benchmarked gems with similar code (e.g. local MemoWise vs. the version pulled from GitHub) will not have their benchmark results depend on whichever happens to run second (which the Ruby interpreter has had more time to optimize). This issue was not seen in CI but was reported by one user running benchmarks locally.

Before merging:

• Copy the table printed at the end of the latest benchmark results into the README.md and update this PR
• If this change merits an update to CHANGELOG.md, add an entry following Keep a Changelog guidelines with semantic versioning

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (ef52a1a) to head (a43489d).

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #355   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            2         2           
  Lines          190       190           
  Branches        90        90           
=========================================
  Hits           190       190           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tagliala]

Does not look to help at all on GitHub actions

2.7

https://github.com/panorama-ed/memo_wise/actions/runs/11040286867/job/30667918311?pr=355#step:11:359

Method arguments	memo_wise-github-main (1.10.0)
() (none)	1.00x
(a)	1.00x
(a, b)	1.03x
(a:)	1.00x
(a:, b:)	1.02x
(a, b:)	1.00x
(a, *args)	1.01x
(a:, **kwargs)	1.00x
(a, *args, b:, **kwargs)	1.00x

3.3

https://github.com/panorama-ed/memo_wise/actions/runs/11040286867/job/30667919858?pr=355#step:11:362

Method arguments	memo_wise-github-main (1.10.0)
() (none)	1.00x
(a)	1.06x
(a, b)	1.00x
(a:)	0.93x
(a:, b:)	0.99x
(a, b:)	1.00x
(a, *args)	0.97x
(a:, **kwargs)	1.00x
(a, *args, b:, **kwargs)	0.98x

---

A previous run without changes to production code lead these results for 3.3:

https://github.com/panorama-ed/memo_wise/actions/runs/10908331323/job/30274052702#step:11:253

Method arguments	memo_wise-github-main (1.10.0)
() (none)	1.00x
(a)	1.06x
(a, b)	1.00x
(a:)	0.93x
(a:, b:)	0.99x
(a, b:)	1.00x
(a, *args)	0.97x
(a:, **kwargs)	1.00x
(a, *args, b:, **kwargs)	0.98x

[pboling]

The fact that it is not helped at all on GHA may be due to resource constraints. The memory & CPU limitations will reduce the amount of optimizations that can be front-loaded.

I don't think that is avoidable except to run one benchmark per action, and collate the results after... :(