rubyperf Ruby performance guide (v0)rubyperf rule ids (slugs)rubyperf corpus targets rubyperf evidence policy

rb-array-plus-rebind — a = a + [x] vs <<rb-file-read-hot — measured, still hint rb-json-hot — measured, still hint rb-puts-hot — puts each vs join once rb-regexp-new-hot — measured, still hint rb-string-plus-hot — string += vs <<

rb-date-parse-in-loop rb-date-parse-in-loop — removed 2026-10-23 (Asia/Shanghai)rb-file-read-in-loop — renamed 2026-10-27 (Asia/Shanghai)rb-hash-merge-in-loop — Hash#merge in loop vs build-once rb-hash-merge-in-loop — removed 2026-10-23 (Asia/Shanghai)rb-json-in-loop — renamed 2026-10-27 (Asia/Shanghai)rb-puts-in-loop — renamed 2026-10-27 (Asia/Shanghai)rb-regexp-new-in-loop — renamed 2026-10-27 (Asia/Shanghai)rb-string-plus-in-loop — renamed 2026-10-27 (Asia/Shanghai)

Project

Contributing Changelog Security Policy Contributor Covenant Code of Conduct AGENT.md — XPerf agent operating contract

LanguagesRubyEvidence_removed

rb-hash-merge-in-loop — removed 2026-10-23 (Asia/Shanghai)

Status: removed. On Ruby 4.0.8 the suggested Hash literal was slower than Hash#merge (0.784×, 7.947 ms vs 6.228 ms, N=50000) and allocated about twice as many objects (200001 vs 100005). A second pass the same evening, load average about 24, still had merge faster (0.645× and 0.655×).

Slug never reused.
No active matcher.

rb-hash-merge-in-loop — Hash#merge in loop vs build-once

Previous Page

rb-json-in-loop — renamed 2026-10-27 (Asia/Shanghai)

Next Page