From 9e1c414595d9096ea113f2794359478e4ca37935 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Bern=C3=A1t=20G=C3=A1bor?= Date: Sun, 12 Jul 2026 16:11:25 -0700 Subject: [PATCH] =?UTF-8?q?=F0=9F=91=B7=20ci(perf):=20measure=20CodSpeed?= =?UTF-8?q?=20with=20macro=20walltime?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit CPU simulation produced incompatible regression sets for unchanged OCI and PyPI benchmark code. PR #213 changed one search test between measured heads; the reported PyPI regressions changed names and magnitudes. Run the mixed microbenchmark and end-to-end operation suites with walltime on the existing pinned macro runner. CodSpeed recommends this mode for I/O and system-level work; repeated sampling avoids one-shot cache and allocator swings. Fixes #221 --- .github/workflows/codspeed.yml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/codspeed.yml b/.github/workflows/codspeed.yml index 307ef17b..ed4ac47f 100644 --- a/.github/workflows/codspeed.yml +++ b/.github/workflows/codspeed.yml @@ -31,8 +31,8 @@ jobs: - uses: taiki-e/install-action@16b05812d776ae1dfaabc8277e421fb6d2506419 # v2.82.7 with: tool: cargo-codspeed - - run: cargo codspeed build -m simulation -p peryx-ecosystem-pypi -p peryx-ecosystem-oci --bench operations + - run: cargo codspeed build -m walltime -p peryx-ecosystem-pypi -p peryx-ecosystem-oci --bench operations - uses: CodSpeedHQ/action@a4a36bb07c0638b0b4ca52bf1f3dad1b4289e52f # v4.18.1 with: - mode: simulation + mode: walltime run: cargo codspeed run