ODS V3 Pre-Release: Public testing and refinement ahead of the official V3 launch. Turn your PC, Mac, or Linux box into a private AI server.
Python ai-agentsamdcomfyuidockerllama-cppllmlocal-ain8nnvidiaopen-webuiragself-hostedspeech-to-textstrix-halotext-to-speechworkflow-automation
"""Exercise NVIDIA's current MIG CSV contract at the topology boundary."""
import json
import os
import subprocess
from pathlib import Path
import pytest
ROOT = Path(__file__).resolve().parents[1]
LIBRARY = ROOT / "installers/lib/nvidia-topo.sh"
@pytest.mark.parametrize(
("current", "pending", "expected"),
[
("Enabled", "Disabled", True),
("Disabled", "Enabled", False),
("Disabled\nEnabled\n[N/A]", "Disabled", True),
("[N/A]\nDisabled", "Enabled", False),
(" Enabled \r\n", "Disabled", True),
],
)
def test_topology_uses_current_mig_state(tmp_path, current, pending, expected):
stub = tmp_path / "nvidia-smi"
stub.write_text(
"""#!/usr/bin/env bash
case "$*" in
--query-gpu=index,name,*) echo '0, NVIDIA A100, 81920, 81920, 4, 16, GPU-one' ;;
--query-gpu=mig.mode.current*) printf '%s\\n' "$MIG_CURRENT" ;;
--query-gpu=driver_version*) echo '580.0' ;;
'topo -m') printf ' GPU0\\nGPU0 X\\n' ;;
-q)
while IFS= read -r current; do
printf ' MIG Mode\\n Current : %s\\n Pending : %s\\n' "$current" "$MIG_PENDING"
done <<< "$MIG_CURRENT"
;;
*) exit 1 ;;
esac
"""
)
stub.chmod(0o755)
result = subprocess.run(
["bash", "-c", 'set -euo pipefail; source "$1"; detect_nvidia_topo', "test", str(LIBRARY)],
env={**os.environ, "PATH": f"{tmp_path}:{os.environ['PATH']}",
"MIG_CURRENT": current, "MIG_PENDING": pending},
capture_output=True, text=True, check=True, timeout=10,
)
topology = json.loads(result.stdout)
assert topology["mig_enabled"] is expected
assert topology["gpus"][0]["uuid"] == "GPU-one"