Skip to content

Install Ipopt Manually

Manual compilation is useful when a platform has no suitable binary or when you need a particular BLAS/LAPACK implementation or sparse linear solver. It is not required for most users; try the Conda-forge procedure in Quick Installation first.

Steps

  1. Build and install Ipopt by following the official Ipopt installation guide. Record the library and header locations.
  2. Make the installed library discoverable. On Unix-like systems this commonly means configuring PKG_CONFIG_PATH and the runtime library path; on Windows it means making the Ipopt DLL directory available on PATH.
  3. Create a fresh Python 3.11 or later environment.
  4. Build cyipopt against that Ipopt installation, then install pockit:
python -m pip install --upgrade pip setuptools wheel
python -m pip install "cyipopt>=1.3.0"
python -m pip install "pockit-optimal-control[examples]"
python -c "import cyipopt; from pockit.optimizer import ipopt; print(cyipopt.__version__)"

cyipopt's build output should identify the intended Ipopt installation. A successful import cyipopt is necessary but not sufficient: run a small pockit example to verify that the linked linear solver also works.

For reproducible deployments, record the Ipopt, cyipopt, BLAS/LAPACK, and sparse linear-solver versions together.