👷 ci(perf): measure CodSpeed with macro walltime

[gaborbernat]

CodSpeed produced incompatible regression sets for PR #213 although the benchmarked OCI and PyPI production sources did not change; one run reported six regressions up to 99.79%, while two reruns replaced them with different sets. Compare all three measurements in #221.

CodSpeed documents cache and allocator variance when CPU simulation runs multiple benchmarks in one binary, and recommends macro-runner walltime for system and I/O suites. The workflow now builds and runs both operation suites in walltime mode while retaining pinned codspeed-macro hardware.

Walltime repeats each case and includes file-system behavior used by the end-to-end HTTP benchmarks. Performance regressions still fail the required CodSpeed analysis.

[read-the-docs-community]

Documentation build overview

📚 peryx | 🛠️ Build #33555616 | 📁 Comparing 9e1c414 against latest (d29b7fe)

  🔍 Preview build  

128 files changed · ± 128 modified

± Modified

• brand-book.html
• index.html
• brand/index.html
• contributing/index.html
• core/index.html
• ecosystems/index.html
• contributing/architecture/index.html
• contributing/benchmarking/index.html
• contributing/release/index.html
• core/access-explained/index.html
• and 118 more...

[codspeed-hq]

Merging this PR will create unknown performance changes

🆕 29 new benchmarks
⏩ 62 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
🆕	WallTime	oci_blob_serve	N/A	161 µs	N/A
🆕	WallTime	oci_manifest_by_digest	N/A	23.6 µs	N/A
🆕	WallTime	oci_tags_list	N/A	18.6 µs	N/A
🆕	WallTime	oci_version_check	N/A	5.5 µs	N/A
🆕	WallTime	normalize_name	N/A	267 ns	N/A
🆕	WallTime	parse_version	N/A	655 ns	N/A
🆕	WallTime	parse_version_specifiers	N/A	678 ns	N/A
🆕	WallTime	sorted_desc	N/A	178.6 µs	N/A
🆕	WallTime	detail_html[large]	N/A	4.6 ms	N/A
🆕	WallTime	detail_html[small]	N/A	37.8 µs	N/A
🆕	WallTime	detail_json[large]	N/A	1.1 ms	N/A
🆕	WallTime	detail_json[small]	N/A	6.8 µs	N/A
🆕	WallTime	distribution_filename	N/A	843 ns	N/A
🆕	WallTime	index_html	N/A	298.2 µs	N/A
🆕	WallTime	index_json	N/A	117.9 µs	N/A
🆕	WallTime	metadata	N/A	6 µs	N/A
🆕	WallTime	detail_html[large]	N/A	305.2 µs	N/A
🆕	WallTime	detail_html[small]	N/A	2.3 µs	N/A
🆕	WallTime	index_html	N/A	56.7 µs	N/A
🆕	WallTime	legacy_json[large]	N/A	7.6 ms	N/A
...	...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

_{Comparing gaborbernat:fix/codspeed-walltime (9e1c414) with main (d29b7fe)}

Footnotes

1. 62 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩