pelagos-decoder 0.1.0__tar.gz

pelagos_decoder-0.1.0/.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,42 @@
+name: Bug report
+description: Something does not work as documented, or gives wrong results.
+labels: ["bug"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thanks for reporting! Please report security vulnerabilities privately; see SECURITY.md.
+  - type: textarea
+    id: what-happened
+    attributes:
+      label: What happened?
+      description: What you expected, what happened instead, and the full error message if there is one.
+    validations:
+      required: true
+  - type: textarea
+    id: reproducer
+    attributes:
+      label: Minimal reproducer
+      description: The smallest code or command that shows the problem.
+      render: python
+    validations:
+      required: true
+  - type: dropdown
+    id: backend
+    attributes:
+      label: GPU backend
+      options:
+        - CUDA (NVIDIA)
+        - ROCm (AMD)
+        - Other / not applicable
+    validations:
+      required: true
+  - type: textarea
+    id: environment
+    attributes:
+      label: Environment
+      description: Paste the output of the command below.
+      value: |
+        python -c "import pelagos, torch, triton, platform; print(platform.platform()); print('python', platform.python_version()); print('torch', torch.__version__, 'cuda', torch.version.cuda, 'hip', torch.version.hip); print('triton', triton.__version__); print('gpu', torch.cuda.get_device_name(0) if torch.cuda.is_available() else None)"
+    validations:
+      required: true

pelagos_decoder-0.1.0/.github/ISSUE_TEMPLATE/config.yml

@@ -0,0 +1,5 @@
+blank_issues_enabled: false
+contact_links:
+  - name: Report a security vulnerability
+    url: https://github.com/quantum-supercomputing/pelagos/security/advisories/new
+    about: Please report vulnerabilities privately, not in a public issue.

pelagos_decoder-0.1.0/.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,19 @@
+name: Feature request
+description: Suggest a new feature or an improvement.
+labels: ["enhancement"]
+body:
+  - type: textarea
+    id: problem
+    attributes:
+      label: What problem would this solve?
+    validations:
+      required: true
+  - type: textarea
+    id: proposal
+    attributes:
+      label: Proposed solution
+      description: How you imagine it working, including any API sketch.
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives considered

pelagos_decoder-0.1.0/.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,12 @@
+## Summary
+
+<!-- What does this change, and why? Link related issues, e.g. "Fixes #123". -->
+
+## Checklist
+
+- [ ] Commits are signed off (`git commit -s`); see [CONTRIBUTING.md](../CONTRIBUTING.md).
+- [ ] Tests added or updated where it makes sense.
+- [ ] `uv run pytest tests/ -v -rs` passes locally.
+- [ ] **If kernels, decoder, or adapter changed:** full suite run on a GPU with no skips.
+      GPU used: <!-- e.g. NVIDIA A100 / AMD MI300X, or "none, maintainer please run" -->
+- [ ] Tutorial notebooks re-run and committed with outputs (if changed).

pelagos_decoder-0.1.0/.github/dependabot.yml

@@ -0,0 +1,8 @@
+version: 2
+updates:
+  # Keep GitHub Actions versions up to date. Python dependencies are pinned in uv.lock and
+  # updated by hand, since torch/triton upgrades need a GPU test run.
+  - package-ecosystem: github-actions
+    directory: /
+    schedule:
+      interval: monthly

pelagos_decoder-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,57 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  cpu:
+    name: CPU checks (Python ${{ matrix.python-version }})
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: ["3.11", "3.12", "3.13"]
+    steps:
+      - uses: actions/checkout@v7
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v7
+        with:
+          enable-cache: true
+
+      # `import pelagos` pulls in the Triton kernels, which import torch at
+      # module scope, so the import-surface tests need the `gpu` extra even on a
+      # CPU runner. torch installs fine without a GPU; the decode tests then
+      # skip on `torch.cuda.is_available()`.
+      # The stim extra lets the DEM -> check-matrix conversion be tested on CPU. The rust
+      # extra is left out: relay-bp only publishes wheels for some Python versions.
+      - name: Sync (gpu extra for torch; no GPU hardware needed)
+        run: uv sync --extra gpu --extra stim --extra test
+
+      # Syntax errors and undefined names only: the lint rules that catch real bugs.
+      - name: Lint
+        run: uvx ruff check --isolated --select E9,F63,F7,F82 src tests tutorials pelagos_benchmark.py
+
+      # Decode tests and the Rust comparison skip without a GPU; -rs lists what was skipped.
+      - name: Tests (CPU)
+        run: uv run pytest tests/ -v -rs
+
+  build:
+    name: Build wheel
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+      - uses: astral-sh/setup-uv@v7
+      - run: uv build
+      - run: ls -la dist/
+
+# The GPU lane is intentionally absent here: Triton kernels need a CUDA or
+# ROCm runtime, which the free GitHub-hosted runners do not provide. The
+# numerical gates below are run by a maintainer on a GPU before each release
+# (see "Testing" in README.md), with
+# `uv sync --extra gpu --extra rust --extra stim --extra test`:
+#
+#   uv run pytest tests/test_triton_against_rust.py -v   # bit-exactness vs. the Rust reference
+#   uv run pytest tests/ -q                              # full suite

pelagos_decoder-0.1.0/.github/workflows/release.yml

@@ -0,0 +1,46 @@
+name: Release
+
+# Publishes to PyPI when a GitHub release is published (tag it vX.Y.Z to match the version
+# in pyproject.toml). Uses PyPI trusted publishing, so no API token is stored in the repo.
+# One-time setup: on PyPI, add this repository and workflow (release.yml) as a trusted
+# publisher for the "pelagos-decoder" project, using the "pypi" environment.
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  build:
+    name: Build distributions
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v7
+      - uses: astral-sh/setup-uv@v7
+
+      - name: Check the tag matches the package version
+        run: |
+          version=$(grep -m1 '^version = ' pyproject.toml | cut -d'"' -f2)
+          if [ "${GITHUB_REF_NAME}" != "v${version}" ]; then
+            echo "Tag ${GITHUB_REF_NAME} does not match pyproject.toml version v${version}"
+            exit 1
+          fi
+
+      - run: uv build
+      - uses: actions/upload-artifact@v7
+        with:
+          name: dist
+          path: dist/
+
+  publish:
+    name: Publish to PyPI
+    needs: build
+    runs-on: ubuntu-latest
+    environment: pypi
+    permissions:
+      id-token: write   # required for trusted publishing
+    steps:
+      - uses: actions/download-artifact@v8
+        with:
+          name: dist
+          path: dist/
+      - uses: pypa/gh-action-pypi-publish@release/v1

pelagos_decoder-0.1.0/.gitignore

@@ -0,0 +1,20 @@
+__pycache__/
+*.py[cod]
+*.so
+.venv/
+venv/
+build/
+dist/
+*.egg-info/
+.pytest_cache/
+.ruff_cache/
+.mypy_cache/
+*.ncu-rep
+*.nsys-rep
+*.sqlite
+autotuner_*.csv
+
+
+# Editor and notebook clutter
+.ipynb_checkpoints/
+.~lock.*#

pelagos_decoder-0.1.0/CHANGELOG.md

@@ -0,0 +1,49 @@
+<!-- Modified 2026 by AMD to document process-per-GPU execution. -->
+
+# Changelog
+
+All notable changes to Pelagos are documented in this file. The format is based on
+[Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and the project follows
+[Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Added
+
+- Opt-in process-per-GPU execution (`backend_config={"execution": "process"}` in
+  `RelayDecoder`, or `pelagos.process.ProcessRelayBPDecoder`), with one persistent worker per
+  GPU, parallel host-side preparation of syndromes and observables, and decoded errors
+  returned through shared memory. The default backend and the decoding kernels are unchanged.
+  See [docs/process-backend.md](docs/process-backend.md).
+
+## [0.1.0] - Unreleased
+
+First public release.
+
+### Added
+
+- Relay-BP-S and plain min-sum BP decoders for qLDPC codes, as fused Triton kernels
+  (`RelayBPDecoder`), batched across syndromes, with `float16`, `float32` and `float64` data types.
+- CUDA and ROCm support through Triton; the backend is detected at construction time.
+- Batch-independent results: each shot stops on its own `stop_nconv`, and the `gamma_draw`
+  option (default `"broadcast"`) shares each relay leg's memory strengths across the batch, so
+  a shot decodes the same regardless of batch size or position. `gamma_draw="per_row"` draws
+  per shot instead.
+- Iteration counts and explicit-gamma rows that match the `relay-bp` Rust reference, and a
+  `valid_mask` that is True exactly where `found_count >= stop_nconv`.
+- `unconverged_correction` option for shots that find no solution: `"first_leg"` (default,
+  as in the Rust reference) returns the hard decision at the end of the pre-iter leg, or of
+  relay leg 1 when `pre_iter=0`; `"last_leg"` returns the one from the last leg that ran.
+- `DecodeResult.found_count`: the number of solutions found for each shot, reported by both
+  `RelayDecoder` detailed-batch methods.
+- `RelayDecoder` and `ObservableDecoderRunner` adapters with an interface compatible with the
+  `relay-bp` Rust package.
+- `CheckMatrices.from_dem` for building decoding problems from stim detector error models.
+- `pelagos_benchmark.py` benchmark CLI, with a Rust backend for side-by-side comparison.
+- Bundled stim circuits for the [[18,4,3]], [[72,12,6]], [[144,12,12]] and [[288,12,18]]
+  bivariate bicycle codes.
+- Tutorial 1: a single-GPU quickstart notebook.
+- Tests, including bit-exact comparison against the Rust reference (GPU), and CPU-only CI.
+
+[Unreleased]: https://github.com/quantum-supercomputing/pelagos/compare/v0.1.0...HEAD
+[0.1.0]: https://github.com/quantum-supercomputing/pelagos/releases/tag/v0.1.0

pelagos_decoder-0.1.0/CITATION.cff

@@ -0,0 +1,90 @@
+cff-version: 1.2.0
+message: "If you use Pelagos in your research, please cite the paper below."
+title: "Pelagos: GPU-accelerated Relay-BP decoder for qLDPC codes"
+type: software
+license: Apache-2.0
+version: 0.1.0
+repository-code: "https://github.com/quantum-supercomputing/pelagos"
+keywords:
+  - quantum error correction
+  - qLDPC codes
+  - belief propagation
+  - Relay-BP
+  - GPU
+  - Triton
+authors:
+  - family-names: "Diamantopoulos"
+    given-names: "D."
+  - family-names: "von Niederhäusern"
+    given-names: "L."
+  - family-names: "Anghel"
+    given-names: "A. S."
+  - family-names: "Kayi"
+    given-names: "A."
+  - family-names: "Pletka"
+    given-names: "R."
+  - family-names: "Müller"
+    given-names: "T."
+  - family-names: "Papandreou"
+    given-names: "N."
+  - family-names: "Blanusa"
+    given-names: "J."
+  - family-names: "Sarafijanovic"
+    given-names: "S."
+  - family-names: "Gonzalez"
+    given-names: "N. M."
+  - family-names: "Abel"
+    given-names: "F."
+  - family-names: "Hagleitner"
+    given-names: "C."
+  - family-names: "Seelam"
+    given-names: "S."
+  - family-names: "Johnson"
+    given-names: "B."
+  - family-names: "Haas"
+    given-names: "R."
+  - family-names: "Pozidis"
+    given-names: "H."
+preferred-citation:
+  type: conference-paper
+  title: "Pêlagos: An Adaptive GPU Decoder for Error Correction Simulations of Quantum LDPC Codes"
+  collection-title: "2026 IEEE International Conference on Quantum Software (QSW)"
+  publisher:
+    name: "IEEE"
+  year: 2026
+  start: 152
+  end: 165
+  url: "https://ieeexplore.ieee.org/document/11662198"
+  authors:
+    - family-names: "Diamantopoulos"
+      given-names: "D."
+    - family-names: "von Niederhäusern"
+      given-names: "L."
+    - family-names: "Anghel"
+      given-names: "A. S."
+    - family-names: "Kayi"
+      given-names: "A."
+    - family-names: "Pletka"
+      given-names: "R."
+    - family-names: "Müller"
+      given-names: "T."
+    - family-names: "Papandreou"
+      given-names: "N."
+    - family-names: "Blanusa"
+      given-names: "J."
+    - family-names: "Sarafijanovic"
+      given-names: "S."
+    - family-names: "Gonzalez"
+      given-names: "N. M."
+    - family-names: "Abel"
+      given-names: "F."
+    - family-names: "Hagleitner"
+      given-names: "C."
+    - family-names: "Seelam"
+      given-names: "S."
+    - family-names: "Johnson"
+      given-names: "B."
+    - family-names: "Haas"
+      given-names: "R."
+    - family-names: "Pozidis"
+      given-names: "H."

pelagos_decoder-0.1.0/CODE_OF_CONDUCT.md

@@ -0,0 +1,84 @@
+
+# Contributor Covenant Code of Conduct
+
+## Our Pledge
+
+We as members, contributors, and leaders pledge to make participation in our community a harassment-free experience for everyone, regardless of age, body size, visible or invisible disability, ethnicity, sex characteristics, gender identity and expression, level of experience, education, socio-economic status, nationality, personal appearance, race, caste, color, religion, or sexual identity and orientation.
+
+We pledge to act and interact in ways that contribute to an open, welcoming, diverse, inclusive, and healthy community.
+
+## Our Standards
+
+Examples of behavior that contributes to a positive environment for our community include:
+
+* Demonstrating empathy and kindness toward other people
+* Being respectful of differing opinions, viewpoints, and experiences
+* Giving and gracefully accepting constructive feedback
+* Accepting responsibility and apologizing to those affected by our mistakes, and learning from the experience
+* Focusing on what is best not just for us as individuals, but for the overall community
+
+Examples of unacceptable behavior include:
+
+* The use of sexualized language or imagery, and sexual attention or advances of any kind
+* Trolling, insulting or derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or email address, without their explicit permission
+* Other conduct which could reasonably be considered inappropriate in a professional setting
+
+## Enforcement Responsibilities
+
+Community leaders are responsible for clarifying and enforcing our standards of acceptable behavior and will take appropriate and fair corrective action in response to any behavior that they deem inappropriate, threatening, offensive, or harmful.
+
+Community leaders have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, and will communicate reasons for moderation decisions when appropriate.
+
+## Scope
+
+This Code of Conduct applies within all community spaces, and also applies when an individual is officially representing the community in public spaces. Examples of representing our community include using an official e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be reported to the community leaders responsible for enforcement by contacting any of the active maintainers listed in [MAINTAINERS.md](MAINTAINERS.md). All complaints will be reviewed and investigated promptly and fairly.
+
+All community leaders are obligated to respect the privacy and security of the reporter of any incident.
+
+## Enforcement Guidelines
+
+Community leaders will follow these Community Impact Guidelines in determining the consequences for any action they deem in violation of this Code of Conduct:
+
+### 1. Correction
+
+**Community Impact**: Use of inappropriate language or other behavior deemed unprofessional or unwelcome in the community.
+
+**Consequence**: A private, written warning from community leaders, providing clarity around the nature of the violation and an explanation of why the behavior was inappropriate. A public apology may be requested.
+
+### 2. Warning
+
+**Community Impact**: A violation through a single incident or series of actions.
+
+**Consequence**: A warning with consequences for continued behavior. No interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, for a specified period of time. This includes avoiding interactions in community spaces as well as external channels like social media. Violating these terms may lead to a temporary or permanent ban.
+
+### 3. Temporary Ban
+
+**Community Impact**: A serious violation of community standards, including sustained inappropriate behavior.
+
+**Consequence**: A temporary ban from any sort of interaction or public communication with the community for a specified period of time. No public or private interaction with the people involved, including unsolicited interaction with those enforcing the Code of Conduct, is allowed during this period. Violating these terms may lead to a permanent ban.
+
+### 4. Permanent Ban
+
+**Community Impact**: Demonstrating a pattern of violation of community standards, including sustained inappropriate behavior, harassment of an individual, or aggression toward or disparagement of classes of individuals.
+
+**Consequence**: A permanent ban from any sort of public interaction within the community.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 2.1, available at [https://www.contributor-covenant.org/version/2/1/code_of_conduct.html][v2.1].
+
+Community Impact Guidelines were inspired by [Mozilla's code of conduct enforcement ladder][Mozilla CoC].
+
+For answers to common questions about this code of conduct, see the FAQ at [https://www.contributor-covenant.org/faq][FAQ]. Translations are available at [https://www.contributor-covenant.org/translations][translations].
+
+[homepage]: https://www.contributor-covenant.org
+[v2.1]: https://www.contributor-covenant.org/version/2/1/code_of_conduct.html
+[Mozilla CoC]: https://github.com/mozilla/diversity
+[FAQ]: https://www.contributor-covenant.org/faq
+[translations]: https://www.contributor-covenant.org/translations

pelagos_decoder-0.1.0/CONTRIBUTING.md

@@ -0,0 +1,69 @@
+# Contributing to Pelagos
+
+Thank you for your interest in Pelagos! Bug reports, fixes, new tutorials, and performance
+improvements are all welcome. By taking part you agree to follow our
+[Code of Conduct](CODE_OF_CONDUCT.md). How decisions are made is described in
+[GOVERNANCE.md](GOVERNANCE.md).
+
+## Reporting issues
+
+Please open a [GitHub issue](https://github.com/quantum-supercomputing/pelagos/issues)
+using one of the templates. For a bug, include the GPU model, the driver version, the
+backend (CUDA or ROCm), and a minimal reproducer. Report security vulnerabilities privately,
+as described in [SECURITY.md](SECURITY.md), not in a public issue.
+
+## Development setup
+
+Pelagos uses [uv](https://docs.astral.sh/uv/):
+
+```bash
+git clone https://github.com/quantum-supercomputing/pelagos.git
+cd pelagos
+uv sync --extra all
+```
+
+## Testing
+
+CI runs on GitHub-hosted runners, which have no GPU. It runs the lint step and the CPU tests;
+the decoder tests are skipped there. Before you open a pull request, run the checks CI runs:
+
+```bash
+uvx ruff check --isolated --select E9,F63,F7,F82 src tests tutorials pelagos_benchmark.py
+uv run pytest tests/ -v -rs
+```
+
+If your change touches the kernels, the decoder, or the adapter, also run the full suite on a
+GPU (CUDA or ROCm) with the `gpu`, `rust`, `stim` and `test` extras installed. Nothing should
+be reported as skipped. Say in the pull request which GPU you used. If you do not have a GPU,
+say so, and a maintainer will run it.
+
+If you change a tutorial notebook, re-run it top to bottom and commit it with its outputs.
+
+## Pull requests
+
+* Keep each pull request focused on one change, and add or update tests where possible.
+* Match the style of the surrounding code.
+* New source files carry an Apache 2.0 license header, either the IBM header used throughout
+  the repository or your own organization's copyright line with
+  `SPDX-License-Identifier: Apache-2.0`. When you modify a file, add a line to its header
+  noting the change.
+* Every commit must be signed off (see below).
+* A maintainer reviews every pull request. CI must pass before merging.
+
+## Developer Certificate of Origin
+
+Pelagos uses the [Developer Certificate of Origin](https://developercertificate.org/)
+(DCO). By signing off a commit, you certify that you wrote the change, or otherwise have the
+right to submit it under the project's [Apache 2.0 license](LICENSE). To sign off, add a
+`Signed-off-by` line to each commit message, which `git commit -s` does for you:
+
+```
+Signed-off-by: Jane Doe <jane.doe@example.com>
+```
+
+The name and email must match the commit author. To sign off commits you already made, run
+`git rebase --signoff main` and force-push your branch.
+
+## License
+
+Contributions are licensed under the [Apache License, Version 2.0](LICENSE).

pelagos_decoder-0.1.0/GOVERNANCE.md

@@ -0,0 +1,41 @@
+# Governance
+
+Pelagos uses a lightweight, maintainer-led governance model. The current maintainers are
+listed in [MAINTAINERS.md](MAINTAINERS.md).
+
+## Maintainers
+
+Maintainers are trusted contributors responsible for project direction, releases, and
+contribution review.
+
+An *active maintainer* is anyone listed under "Active maintainers" in
+[MAINTAINERS.md](MAINTAINERS.md). A maintainer may step down at any time by opening a pull
+request that moves their name to "Emeritus maintainers." Emeritus maintainers do not vote,
+and they may return by the same process as a new maintainer.
+
+## Decision-making
+
+Routine decisions, such as reviewing and merging pull requests or triaging issues, are
+delegated to the relevant maintainers.
+
+Significant decisions are discussed publicly in a GitHub issue or pull request and decided
+by maintainer consensus. Significant decisions include changes to the public API, the
+release process, the license, or the project's scope. If consensus cannot be reached, a
+majority of active maintainers decides.
+
+## Membership
+
+New maintainers are nominated on the basis of sustained contributions to the project. A
+nomination is made by opening a pull request that adds the nominee to
+[MAINTAINERS.md](MAINTAINERS.md). The nomination is approved by a majority of active
+maintainers.
+
+## Disagreements
+
+Disputes should first be resolved through good-faith discussion. If a dispute remains
+unresolved, maintainers who are not involved in it make the final decision.
+
+## Changes to governance
+
+Changes to this document require public notice, given as a pull request that stays open for
+at least 14 days, and approval by a two-thirds supermajority of active maintainers.