{ "markdown": "\n# cuOpt - GPU-accelerated Optimization\n\n[![Build Status](https://github.com/NVIDIA/cuopt/actions/workflows/build.yaml/badge.svg)](https://github.com/NVIDIA/cuopt/actions/workflows/build.yaml)\n[![Version](https://img.shields.io/badge/version-26.10.00-blue)](https://github.com/NVIDIA/cuopt/releases)\n[![Documentation](https://img.shields.io/badge/docs-latest-brightgreen)](https://docs.nvidia.com/cuopt/user-guide/latest/introduction.html)\n[![Docker Hub](https://img.shields.io/badge/docker-nvidia%2Fcuopt-blue?logo=docker)](https://hub.docker.com/r/nvidia/cuopt)\n[![Examples](https://img.shields.io/badge/examples-cuopt--examples-orange)](https://github.com/NVIDIA/cuopt-examples)\n[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/NVIDIA/cuopt-examples/blob/cuopt_examples_launcher/cuopt_examples_launcher.ipynb)\n[![NVIDIA Launchable](https://img.shields.io/badge/NVIDIA-Launchable-76b900?logo=nvidia)](https://brev.nvidia.com/launchable/deploy?launchableID=env-2qIG6yjGKDtdMSjXHcuZX12mDNJ)\n[![Videos and Tutorials](https://img.shields.io/badge/Videos_and_Tutorials-red?logo=youtube)](https://docs.nvidia.com/cuopt/user-guide/latest/resources.html#cuopt-examples-and-tutorials-videos)\n\n\n\nNVIDIA® cuOpt™ is a GPU-accelerated optimization engine that excels in linear programming (LP), quadratic programming (QP), and vehicle routing problems (VRP), with support for quadratically constrained quadratic programming (QCQP) (beta), second-order cone programming (SOCP) (beta), and mixed integer programming (MIP) (beta). cuOpt enables near real-time solutions for large-scale LPs with millions of variables and constraints and offers easy integration into existing modeling languages with seamless deployment across hybrid and multi-cloud environments.\n\nThe cuOpt MIP solver is in beta and under active development. The solver currently excels at finding high-quality feasible solutions quickly with GPU-accelerated primal heuristics. Proving feasible solutions optimal remains under active development.\n\ncuOpt offers easy integration into existing solvers, and seamlessly extends into agent-first optimization workflows through open-source cuOpt agent skills.\n\nThe core engine is written in C++ and wrapped with a C API, Python API and Server API.\n\nFor the latest version, ensure you are on the `main` branch.\n\n## Latest Documentation\n\n[cuOpt Documentation](https://docs.nvidia.com/cuopt/user-guide/latest/introduction.html)\n\n## Supported APIs\n\ncuOpt supports the following APIs:\n\n- C API support\n - Linear Programming (LP)\n - Quadratic Programming (QP)\n - Quadratically Constrained Quadratic Programming (QCQP) (beta)\n - Second-Order Cone Programming (SOCP) (beta)\n - Mixed Integer Programming (MIP) (beta)\n- C++ API support\n - cuOpt is written in C++ and includes a native C++ API. However, we do not provide documentation for the C++ API at this time. We anticipate that the C++ API will change significantly in the future. Use it at your own risk.\n- Python support\n - Routing (TSP, VRP, and PDP)\n - Linear Programming (LP)\n - Quadratic Programming (QP)\n - Quadratically Constrained Quadratic Programming (QCQP) (beta)\n - Second-Order Cone Programming (SOCP) (beta)\n - Mixed Integer Programming (MIP) (beta)\n - Algebraic modeling Python API allows users to easily build constraints and objectives\n- Server support\n - Linear Programming (LP)\n - Routing (TSP, VRP, and PDP)\n - Mixed Integer Programming (MIP) (beta)\n\nThis repo is also hosted as a [COIN-OR](http://github.com/coin-or/cuopt/) project.\n\n## Latest Release Notes:\n\n[RELEASE-NOTES.md](RELEASE-NOTES.md)\n\n## Installation\n\n### CUDA/GPU requirements\n\n* CUDA 12.0+ or CUDA 13.0+\n* NVIDIA driver >= 525.60.13 (Linux) and >= 527.41 (Windows)\n* Volta architecture or better (Compute Capability >=7.0)\n\n### Python requirements\n\n* Python >=3.11, <=3.14\n\n### OS requirements\n\n* Only Linux is supported and Windows via WSL2\n * x86_64 (64-bit)\n * aarch64 (64-bit)\n\nNote: WSL2 is tested to run cuOpt, but not for building.\n\nMore details on system requirements can be found [here](https://docs.nvidia.com/cuopt/user-guide/latest/system-requirements.html)\n\n### Pip\n\nPip wheels are easy to install and easy to configure. Users with existing workflows who uses pip as base to build their workflows can use pip to install cuOpt.\n\ncuOpt can be installed via `pip` from the NVIDIA Python Package Index.\nBe sure to select the appropriate cuOpt package depending\non the major version of CUDA available in your environment:\n\nFor CUDA 12.x:\n\n```bash\npip install \\\n --extra-index-url=https://pypi.nvidia.com \\\n nvidia-cuda-runtime-cu12==12.9.* \\\n cuopt-server-cu12==26.10.* cuopt-sh-client==26.10.*\n```\n\nDevelopment wheels are available as nightlies, please update `--extra-index-url` to `https://pypi.anaconda.org/rapidsai-wheels-nightly/simple/` to install latest nightly packages.\n```bash\npip install --pre \\\n --extra-index-url=https://pypi.nvidia.com \\\n --extra-index-url=https://pypi.anaconda.org/rapidsai-wheels-nightly/simple/ \\\n cuopt-server-cu12==26.10.* cuopt-sh-client==26.10.*\n```\n\nFor CUDA 13.x:\n\n```bash\npip install \\\n --extra-index-url=https://pypi.nvidia.com \\\n cuopt-server-cu13==26.10.* cuopt-sh-client==26.10.*\n```\n\nDevelopment wheels are available as nightlies, please update `--extra-index-url` to `https://pypi.anaconda.org/rapidsai-wheels-nightly/simple/` to install latest nightly packages.\n```bash\npip install --pre \\\n --extra-index-url=https://pypi.nvidia.com \\\n --extra-index-url=https://pypi.anaconda.org/rapidsai-wheels-nightly/simple/ \\\n cuopt-server-cu13==26.10.* cuopt-sh-client==26.10.*\n```\n\n\n### Conda\n\ncuOpt can be installed with conda (via [miniforge](https://github.com/conda-forge/miniforge)):\n\nAll other dependencies are installed automatically when `cuopt-server` and `cuopt-sh-client` are installed.\n\n```bash\nconda install -c rapidsai -c conda-forge -c nvidia cuopt-server=26.10.* cuopt-sh-client=26.10.*\n```\n\nWe also provide [nightly conda packages](https://anaconda.org/rapidsai-nightly) built from the HEAD\nof our latest development branch. Just replace `-c rapidsai` with `-c rapidsai-nightly`.\n\n### Container\n\ncuOpt ships two container variants on Docker Hub — an Ubuntu-based image for general use and a Red Hat Universal Base Image 10 (UBI10) image for FIPS 140-3 compliant environments.\n\n#### Ubuntu image\n\n```bash\n# For CUDA 12.x\ndocker pull nvidia/cuopt:latest-cu12\n\n# For CUDA 13.x\ndocker pull nvidia/cuopt:latest-cu13\n```\n\nThe `latest` tag is the latest stable release. To pin a specific version use `-cu12` / `-cu13` (e.g. `26.6.0-cu13`). Fully-qualified tags that also pin the CUDA minor and Python version (e.g. `26.6.0-cuda13.3-py3.14`) are published alongside the short tags. See the [cuOpt Docker Hub page](https://hub.docker.com/r/nvidia/cuopt/tags) for the full list.\n\n#### UBI10 image (FIPS 140-3)\n\nBased on Red Hat Universal Base Image 10 (RHEL 10), which ships OpenSSL 3.5 validated under FIPS 140-3. Use this image in environments with strict FIPS or RHEL compliance requirements. UBI10 images are available for CUDA 13.x only.\n\n```bash\ndocker pull nvidia/cuopt:latest-cu13-ubi10\n```\n\nFully-qualified tags follow the pattern `-cuda-ubi10` (e.g. `26.6.0-cuda13.3-ubi10`). Nightly builds use the `a-cu13-ubi10` tag scheme. See the [cuOpt Docker Hub page](https://hub.docker.com/r/nvidia/cuopt/tags) for the full list.\n\nBoth images include the same cuOpt packages (`libcuopt`, `cuopt`, `cuopt-server`, `cuopt-sh-client`) and expose the same server entrypoint. They are built and tested for x86-64 and ARM64.\n\nNightly container images for both variants are built from the HEAD of the development branch. They are tagged as `a-cu12` / `a-cu13` (Ubuntu) and `a-cu13-ubi10` (UBI10).\n\nMore information about the cuOpt container can be found [here](https://docs.nvidia.com/cuopt/user-guide/latest/cuopt-server/quick-start.html#container-from-docker-hub).\n\nUsers who are using cuOpt for quick testing or research can use the cuOpt container. Alternatively, users who are planning to plug cuOpt as a service in their workflow can quickly start with the cuOpt container. But users are required to build security layers around the service to safeguard the service from untrusted users.\n\n## Build from Source and Test\n\nPlease see our [guide for building cuOpt from source](CONTRIBUTING.md#setting-up-your-build-environment). This will be helpful if users want to add new features or fix bugs for cuOpt. This would also be very helpful in case users want to customize cuOpt for their own use cases which require changes to the cuOpt source code.\n\n## Release Timeline\n\ncuOpt follows the RAPIDS release schedule and is part of the **\"others\"** category in the release timeline. The release cycle consists of:\n\n- **Development**: Active feature development and bug fixes targeting `main`\n- **Burn Down**: Focus shifts to stabilization; new features should target the next release\n- **Code Freeze**: Only critical bug fixes allowed; PRs require admin approval\n- **Release**: Final testing, tagging, and official release\n\nFor current release timelines and dates, refer to the [RAPIDS Maintainers Docs](https://docs.rapids.ai/maintainers/).\n\n## For AI Coding Agents\n\nSee [AGENTS.md](./AGENTS.md) for agent-specific guidelines.\n\n## Contributing Guide\n\nReview the [CONTRIBUTING.md](CONTRIBUTING.md) file for information on how to contribute code and issues to the project.\n", "bytes": 9597, "sha": "d6d54d87e6d679fe7c7d316f1cbafbf69e51428ecc1305527f0749b7650bdfec", "repo_slug": "nvidia/cuopt", "fonte": "repo", "truncated": false, "api": "https://agentalog.com/api/listings/plg_nvidia_cuopt_c9202e69/readme" }