Files
marqdo-cli/.github/workflows/performance-linux.yml
T
cflmy d2267e912a
CI / fmt + clippy + test (Linux x86_64) (push) Waiting to run
CI / schema instance validation + consumers (push) Waiting to run
CI / runtime contract (macos-latest) (push) Waiting to run
CI / runtime contract (windows-latest) (push) Waiting to run
Linux strict security lane (G-C) / strict isolation escape matrix (push) Waiting to run
ci/woodpecker/push/woodpecker Pipeline was canceled
perf: add correctness-verified marqdo.perf.v2 sampler
Replace the three-run baseline profile with tools/perf/runner.py: one fixed
release binary, per-child rusage via posix_spawn/wait4, 5 warmups and 20
samples by default, SHA-derived interleaved order, p50/p95/MAD, CPU user/sys
and peak RSS. Every warmup output is normalized and compared with a committed
wire digest before timing; a mismatch is CORRECTNESS_FAIL, never a fast
sample. Short or unapproved-host runs are SMOKE, and baselines from another
platform_id/profile are refused.

Add workloads.json, expected digests for nine existing workloads, a dedicated
fixed-node workflow (manual, self-hosted, requires --cpu-list), and
docs/performance.md. bench.py stays as a compatibility shim. This is step 5 of
the CI/security RFC (B08).
2026-11-21 17:28:05 +08:00

71 lines
2.2 KiB
YAML

name: Linux performance measurements (G4)
# This workflow needs a dedicated, fixed Linux performance node. It is never
# run on GitHub-hosted runners or on TCG/QEMU guests: those numbers are not
# comparable with a native release baseline. A short smoke run is not release
# evidence; runner.py marks it SMOKE unless 5 warmups / 20 samples and
# --approved-node are used.
#
# Trust model: the node carries no untrusted PR code and no repository
# secrets. It must not share hardware with the G3 security VM.
on:
workflow_dispatch:
inputs:
platform_id:
description: "Fixed node identifier, e.g. linux-x86_64-fixed-node-01"
required: true
type: string
cpu_list:
description: "taskset CPU list, e.g. 2,3 (empty = unbound smoke only)"
required: false
default: ""
type: string
baseline_run_id:
description: "Optional previous run id whose artifact is the baseline"
required: false
default: ""
type: string
permissions:
contents: read
jobs:
measure:
name: correctness-verified performance sample
runs-on: [self-hosted, linux, marqdo-perf]
timeout-minutes: 60
steps:
- uses: actions/checkout@v4
with:
persist-credentials: false
- name: Toolchain report
run: |
rustc -vV
uname -a
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor 2>/dev/null || true
cat /proc/loadavg
- name: Measure
shell: bash
env:
PLATFORM_ID: ${{ inputs.platform_id }}
CPU_LIST: ${{ inputs.cpu_list }}
run: |
set -euo pipefail
args=(--warmups 5 --samples 20 --approved-node --platform-id "$PLATFORM_ID" --out perf.json)
if [ -n "$CPU_LIST" ]; then
args+=(--cpu-list "$CPU_LIST")
else
echo "error: a release measurement requires a fixed --cpu-list" >&2
exit 1
fi
python3 tools/perf/runner.py "${args[@]}"
- name: Upload report
if: always()
uses: actions/upload-artifact@v4
with:
name: marqdo-perf-${{ inputs.platform_id }}
path: perf.json
if-no-files-found: warn
retention-days: 90