ScratchV/.github/workflows/w3-preparation.yml at main · ScratchV-Compiler/ScratchV · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
name: W3 preparation
on:
  workflow_dispatch:
  workflow_call:
permissions:
  contents: read
concurrency:
  group: w3-preparation-${{ github.ref }}
  cancel-in-progress: false
jobs:
  preparation:
    runs-on: ubuntu-24.04
    timeout-minutes: 60
    env:
      OMP_NUM_THREADS: "1"
      OPENBLAS_NUM_THREADS: "1"
      MKL_NUM_THREADS: "1"
      PYTHONIOENCODING: utf-8
    steps:
      - uses: actions/checkout@v4
        with:
          persist-credentials: false
          # The ONNX benchmark regression archives a pinned historical baseline.
          fetch-depth: 0
      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
        with:
          python-version: "3.12"
      - name: Record Linux runner resources
        run: |
          set -euo pipefail
          mkdir -p output
          {
            printf 'commit=%s\nrun_id=%s\nattempt=%s\n' "$GITHUB_SHA" "$GITHUB_RUN_ID" "$GITHUB_RUN_ATTEMPT"
            cat /etc/os-release
            uname -a
            lscpu
            free -b
            df -B1 .
          } | tee output/w3-runner-resources.txt
      - name: Install fixed CPU environment and QEMU
        run: |
          set -euo pipefail
          sudo apt-get update
          sudo apt-get install --no-install-recommends -y qemu-system-misc
          python -m pip install torch==2.7.1 --index-url https://download.pytorch.org/whl/cpu
          python -m pip install -r requirements/qwen3-small-probe.txt PyYAML==6.0.3
          python -m pip install -e . ziglang==0.14.1
          python -m pip check
          python -m pip freeze > output/w3-dependencies.txt
          python --version >> output/w3-runner-resources.txt
          qemu-system-riscv64 --version >> output/w3-runner-resources.txt
          python -m ziglang version >> output/w3-runner-resources.txt
          echo "SCRATCHV_CC=$(python -c 'from pathlib import Path; import ziglang; print(Path(ziglang.__file__).parent / "zig")')" >> "$GITHUB_ENV"
      - name: W3 and affected W2/IR regressions
        run: |
          set -euo pipefail
          mkdir -p output
          python -B -X utf8 -m pytest tests/test_w3*.py tests/test_qwen3_medium*.py tests/test_ir_interpreter*.py tests/test_onnx*.py tests/test_qwen3_small_model.py tests/test_qwen3_small_probe.py tests/test_qwen3_small_gate.py tests/test_qwen3_small_riscv_gate.py tests/test_w2_runtime_model.py -q -p no:cacheprovider --basetemp output/pytest-w3 --junitxml output/w3-tests.xml
      - name: Download and verify fixed assets
        run: |
          python - <<'PY'
          from pathlib import Path
          from huggingface_hub import snapshot_download
          from probes.w1_qwen3_export.run import MANIFEST, acquire_model
          from probes.w3_qwen3_subgraphs.assets import verify_source
          source = Path("output/assets/source")
          snapshot_download(repo_id=MANIFEST["model_id"], revision=MANIFEST["revision"],
                            local_dir=source, allow_patterns=["model.safetensors"] + [x["name"] for x in MANIFEST["source_files"]])
          verify_source(source)
          acquire_model(Path("output/assets/onnx"))
          PY
      - name: Five preparation checks (not full-model acceptance)
        id: probe
        run: |
          python -B -X utf8 scripts/run_w3_preparation.py --source-dir output/assets/source --model-dir output/assets/onnx --output-dir output/w3-preparation --cc "$SCRATCHV_CC" --qemu qemu-system-riscv64
      - name: Summarize W3 preparation evidence
        if: always()
        env:
          PROBE_OUTCOME: ${{ steps.probe.outcome }}
        run: python -B -X utf8 scripts/summarize_w3_ci.py --kind preparation --report-dir output/w3-preparation
      - name: Retain bounded selected trace and failure evidence
        if: always()
        run: |
          python -B -X utf8 scripts/collect_w3_ci_evidence.py --source output/w3-preparation --destination output/w3-diagnostics
      - name: Upload preparation reports (30 days)
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: w3-preparation-reports
          if-no-files-found: warn
          retention-days: 30
          path: |
            output/w3-runner-resources.txt
            output/w3-dependencies.txt
            output/w3-tests.xml
            output/w3-preparation/**/report.json
            output/w3-preparation/**/report.md
            output/w3-preparation/**/report.html
            output/w3-preparation/**/trace_schema.json
            output/w3-preparation/logs/*.log
            output/w3-diagnostics/retained-files.json
            output/w3-diagnostics/retained-evidence.json
      - name: Upload selected diagnostic evidence (7 days)
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: w3-preparation-diagnostics
          if-no-files-found: warn
          retention-days: 7
          compression-level: 1
          path: |
            output/w3-diagnostics/