NVIDIA cuOpt#

NVIDIA cuOpt is an open-source GPU-accelerated optimization engine delivering near real-time solutions for complex decision-making challenges. The framework used by the driver supports automatic reformulation for many expression types; the modeling guide can be found here.

Try cuOpt on Google Colab using NVIDIA GPUs for free!

Learn More | Modeling guide | Features guide | Options | Changes

System requirements#

CUDA/GPU requirements#

  • CUDA 12.0+

  • NVIDIA driver >= 525.60.13 (Linux) and >= 527.41 (Windows)

  • Volta architecture or better (Compute Capability >=7.0)

OS requirements#

  • Only Linux is supported and Windows via WSL2

    • x86_64 (64-bit)

    • aarch64 (64-bit)

Container#

Use a container from here is reccomended; the runtime tag is needed, as it includes needed maths libraries. An example of supported container is::

$ docker run -it --rm --gpus all nvidia/cuda:12.9.0-runtime-ubuntu24.04 /bin/bash

How to use it#

How to install:

# Install Python API for AMPL:
$ python -m pip install amplpy --upgrade

# Install AMPL & solver modules:
$ python -m amplpy.modules install cuopt # install cuopt

# Activate your license (e.g., free ampl.com/ce or ampl.com/courses licenses):
$ python -m amplpy.modules activate <your-license-uuid>

How to use:

from amplpy import AMPL
ampl = AMPL()
...
ampl.solve(solver="cuopt", cuopt_options="option1=value1 option2=value2")

Try cuOpt on Google Colab using NVIDIA GPUs for free!

Learn more about what we have to offer to implement and deploy Optimization in Python.

AMPL APIs are interfaces that allow developers to access the features of the AMPL interpreter from within a programming language. We have APIs available for:

ampl: option solver cuopt; # change the solver
ampl: option cuopt_options 'option1=value1 option2=value2'; # specify options
ampl: solve; # solve the problem

Resources#

Solver options#

Full list of solver options:

More details on solver options: Features guide.

Retrieving solutions#

The outcome of the last optimization is stored in the AMPL parameter solve_result_num and the relative message in solve_result.

display solve_result_num, solve_result;

cuOpt solve result codes can be obtained by running cuopt -! or ampl: shell "cuopt -!";:

        0- 99 solved: optimal for an optimization problem,
            feasible for a satisfaction problem
    100-199 solved? solution candidate returned but error likely
        150 solved? MP solution check failed (option sol:chk:fail)
    200-299 infeasible
    300-349 unbounded, feasible solution returned
    350-399 unbounded, no feasible solution returned
    400-449 limit, feasible: stopped, e.g., on iterations or Ctrl-C
    450-469 limit, problem is either infeasible or unbounded.
            Disable dual reductions or run IIS finder for definitive answer.
    470-499 limit, no solution returned
    500-999 failure, no solution returned
        550 failure: numeric issue, no feasible solution

For general information, see MP result codes guide.

Changelog#