Mcore Testing

NVIDIAOfficialCreated by Philip PetrakianDistributed by NVIDIA

Skill
Guidelines
Software development
Apache-2.0
Medium risk

From the author

Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.

— the description by Philip Petrakian

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill provides a structured approach to testing within the Megatron-LM ecosystem. It details the directory layout for unit and functional tests, explains the YAML-based recipe system for CI integration, and outlines procedures for local execution, debugging, and managing test lifecycle, including handling flaky tests and golden value updates.

What it does

Defines the directory structure for unit and functional test suites.
Explains the YAML recipe format used for defining and scoping test workloads.
Provides commands for executing tests locally using torch.distributed.run.
Outlines the process for adding new unit and functional tests.
Describes how to manage flaky tests using pytest markers and scope suffixes.
Details the workflow for downloading and committing golden values for functional tests.

When to use it

When adding new unit or functional tests to the Megatron-LM codebase.
When debugging test failures or reproducing CI issues locally.
When modifying test recipes or updating golden values after code changes.
When needing to temporarily disable a test without deleting its configuration.

Advantages

Ensures consistency between local development and CI environments.
Provides clear patterns for handling flaky tests without blocking pipelines.
Simplifies the process of adding new test cases through standardized YAML recipes.
Offers specific guidance on multi-rank testing and distributed execution.

Usage examples

“How do I run a single unit test file locally?”: The user is provided with the specific torch.distributed.run command structure required to execute a single test file with the necessary GPU configuration.
“How can I temporarily disable a functional test in the CI?”: The user is instructed to suffix the scope in the recipe YAML with -broken to skip the test while preserving the configuration for future use.
“What is the process for updating golden values after a code change?”: The user is guided to run the test, then use the download_golden_values.py script with the appropriate pipeline ID to fetch and commit the new values.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/nvidia/mcore-testing/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip mcore-testing.zip -d .agents/skills

Documentation by Philip Petrakian

Testing Guide

Answer-First Testing Facts

For questions about disabling tests without deleting them:

Functional recipe entries stay in YAML; disable by suffixing scope with -broken, for example scope: [mr-github] -> scope: [mr-github-broken].
Unit-test skips use pytest markers instead: @pytest.mark.flaky_in_dev skips in the default dev environment, and @pytest.mark.flaky skips in LTS.
Do not delete the test case or recipe entry when the goal is discoverability and easy re-enable.

Test Layout

tests/
├── unit_tests/          # pytest, 1 node × 8 GPUs, torch.distributed runner
├── functional_tests/    # end-to-end shell + training scripts
│   └── test_cases/
│       └── {model}/{test_case}/
│           ├── model_config.yaml          # training args
│           └── golden_values_{env}_{platform}.json
└── test_utils/
    ├── recipes/
    │   ├── h100/        # YAML recipes for H100 jobs
    │   └── gb200/       # YAML recipes for GB200 jobs
    └── python_scripts/  # helpers (recipe_parser, golden-value download, …)

How Tests Execute

The GitHub Actions runner invokes launch_nemo_run_workload.py, which uses nemo-run to launch a DockerExecutor container. The repo is bind-mounted at /opt/megatron-lm; training data is mounted at /mnt/artifacts.

Unit tests are dispatched through torch.distributed.run:

Ranks 0 and 3 are tee-d to stdout; all other ranks write only to log files.
Per-rank log files land at {assets_dir}/logs/1/ and are uploaded as a GitHub artifact after the run.

Functional tests are driven by tests/functional_tests/shell_test_utils/run_ci_test.sh. Only rank 0 runs the pytest validation step; training output from all ranks is uploaded as an artifact.

Flaky-failure auto-retry: launch_nemo_run_workload.py retries up to 3 times for known transient patterns (NCCL timeout, ECC error, segfault, HuggingFace connectivity, …) before declaring a genuine failure.

Recipe YAML Structure

Recipes live in tests/test_utils/recipes/ and are parsed by tests/test_utils/python_scripts/recipe_parser.py. Each file expands a cartesian products block into individual workload specs:

type: basic
format_version: 1
maintainers: [mcore]
loggers: [stdout]
spec:
  name: "{test_case}_{environment}_{platforms}"
  model: gpt              # maps to tests/functional_tests/test_cases/{model}/
  build: mcore-pyt-{environment}
  nodes: 1
  gpus: 8
  n_repeat: 5
  platforms: dgx_h100
  time_limit: 1800
  script_setup: |
    ...
  script: |-
    bash tests/functional_tests/shell_test_utils/run_ci_test.sh ...
products:
  - test_case: [my_test]
    products:
      - environment: [dev, lts]
        scope: [mr-github]
        platforms: [dgx_h100]

Key runtime placeholders: {assets_dir}, {artifacts_dir}, {test_case}, {environment}, {platforms}, {n_repeat}.

Disabling a Test Without Deleting It

To temporarily disable a test case in a recipe YAML, suffix its scope value with -broken — do not delete the entry:

# before (test runs in CI)
scope: [mr-github]

# after (test is skipped; entry preserved for easy re-enable)
scope: [mr-github-broken]

Running Unit Tests Locally

All unit tests initialize a torch.distributed group, so every invocation requires GPU access and must go through torch.distributed.run:

# Full suite
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests

# Single file
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests/models/test_gpt_model.py

# Single test
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests/models/test_gpt_model.py::TestGPTModel::test_constructor

# Filter by name substring
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests -k optimizer

Marker filters

# Exclude flaky tests during development
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests -m "not flaky and not flaky_in_dev"

# Include experimental tests
uv run python -m torch.distributed.run --nproc-per-node 8 -m pytest -q \
  tests/unit_tests --experimental

CI parity

Use tests/unit_tests/run_ci_test.sh to reproduce a CI bucket failure exactly. For ad-hoc runs, prefer the direct torch.distributed.run invocations above.

Gotchas

pyproject.toml sets addopts = --durations=15 -s -rA — stdout is not captured (-s), so ranks interleave during multi-rank runs. Override with --capture=fd when debugging a specific rank.
tests/unit_tests/conftest.py looks for test data under /opt/data and attempts a download if missing. Supply it manually or skip data-dependent tests when running outside the canonical container.

Adding a Unit Test

Create tests/unit_tests/<category>/test_<name>.py.
Use fixtures from tests/unit_tests/conftest.py.
Apply markers as needed:
- @pytest.mark.internal — skipped on legacy tag
- @pytest.mark.flaky_in_dev — skipped in dev environment (CI default; use this to disable a flaky test without blocking the standard pipeline)
- @pytest.mark.flaky — skipped in lts environment
- @pytest.mark.experimental — latest tag only
Verify locally (see Running Unit Tests Locally above).
If the test needs a dedicated CI bucket, add an entry to tests/test_utils/recipes/h100/unit-tests.yaml.

Adding a Functional / Integration Test

Create tests/functional_tests/test_cases/<model>/<test_name>/.
Write model_config.yaml with MODEL_ARGS, ENV_VARS, and TEST_TYPE.
Add a YAML recipe under tests/test_utils/recipes/h100/ (and gb200/ if needed). Required fields: scope, environment, platform, n_repeat, time_limit.
Push the PR, add the label "Run functional tests" to trigger a full run.

After a successful run, download golden values:

python tests/test_utils/python_scripts/download_golden_values.py \
  --source github --pipeline-id <run-id>

Commit the downloaded golden values.

Common Pitfalls

Problem	Cause	Fix
Test passes locally but fails in CI	Different environment or data path	Check `DATA_PATH`, `DATA_CACHE_PATH`, and the `environment` tag (`dev` vs `lts`)
Golden value mismatch after a code change	Numerical regression	Download new golden values via `download_golden_values.py` after a clean run
`cicd-integration-tests-gb200` not triggered	GB200 jobs require maintainer status	Ask a maintainer to trigger, or add the `Run functional tests` label

Files

ATTRIBUTION.mdAdded by AutoGineer907 B
BENCHMARK.md2.7 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md7.3 KiB
evals/evals.json3 B
skill-card.md2.5 KiB
skill.oms.sig4.5 KiB

skill.oms.sig

{"mediaType":"application/vnd.dev.sigstore.bundle.v0.3+json","verificationMaterial":{"x509CertificateChain":{"certificates":[{"rawBytes":"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"},{"rawBytes":"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"},{"rawBytes":"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"}]},"tlogEntries":[]},"dsseEnvelope":{"payload":"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","payloadType":"application/vnd.in-toto+json","signatures":[{"sig":"MGYCMQDJ8fHzletLrKa3Nxh/YIAbb/TaY0rfmHMf/vSzoBV62KS/zNNk1D6YNleRmHo0pg8CMQDL3q9Q5JN7Edt08AXfrk5UrWyDoUPm+pEgQDbC5yHx/qwC/vJf+hawIp9Z+8qCrVg=","keyid":""}]}}

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Needs Python, Needs Node.js

Security

What our scanner noticed· Medium risk

Large base64-encoded blobs

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

Philip Petrakian

Publisher

NVIDIA (Official)

Distributed by

NVIDIA

Source

NVIDIA/skills · skills/mcore-testing · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE-APACHE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by Philip Petrakian or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download mcore-testing.zip