From 530fbf50e2bc26e7bc1c0aa29450afd992a2d335 Mon Sep 17 00:00:00 2001 From: garrytan Date: Sat, 31 Oct 2026 18:45:07 +0000 Subject: [PATCH] ci: gate census max-parallel 20 -> 24 for the ci-audit-planted behavior trials (Captain exception, GSTA-204) A 3-trial behavior case needs three distinct slices; every gate slice is full, so its trials open two new slices per census arm (claude-opus 18, codex-sol 6: 24 rows against a cap of 20). Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> --- .github/workflows/evals-periodic.yml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/evals-periodic.yml b/.github/workflows/evals-periodic.yml index 2804069c4..001ab11c7 100644 --- a/.github/workflows/evals-periodic.yml +++ b/.github/workflows/evals-periodic.yml @@ -205,7 +205,7 @@ jobs: set -euo pipefail for arm in $(jq -r '.[]' <<<"$ARMS"); do mapfile -t arm_env < <(bun --no-install run scripts/census-arms.ts --env "$arm") - env "${arm_env[@]}" EVALS_TIER=gate bun run scripts/test-paid-shards.ts --tier gate --emit-plan "/tmp/gate-census-plan/$arm/manifest.json" --slice-budget 420 --jobs 2 --skip-judges --max-parallel 20 + env "${arm_env[@]}" EVALS_TIER=gate bun run scripts/test-paid-shards.ts --tier gate --emit-plan "/tmp/gate-census-plan/$arm/manifest.json" --slice-budget 420 --jobs 2 --skip-judges --max-parallel 24 done # One executor per arm and planned slice, each with its slice's ceiling. @@ -574,7 +574,7 @@ jobs: strategy: # Two file workers per slice, each retaining two in-file case workers. fail-fast: false - max-parallel: 20 + max-parallel: 24 matrix: include: ${{ fromJSON(needs.plan-slices.outputs.gate_matrix) }} steps: