Python

The easiest way to install the python version is using pip:

pip install scs

You can also install directly from source

git clone --recursive https://github.com/bodono/scs-python.git
cd scs-python
python -m pip install .

Apple Accelerate (macOS)

On macOS the Apple Accelerate backend is built and included automatically — no extra install flags are needed. It uses the Accelerate framework’s sparse LDL^T solver, which is optimized for Apple hardware including Apple Silicon. See here for how to select Accelerate when solving.

MKL

If you have MKL, you can install the MKL Pardiso interface using

python -m pip install -Csetup-args=-Dlink_mkl=true .

See here for how to enable MKL when solving. MKL is typically faster than the built-in linear system solver.

The published Linux x86_64 wheels prefer the threaded MKL variant and include the Intel OpenMP runtime (libiomp5). Windows currently falls back to sequential MKL until Intel fixes the threaded pkg-config metadata in its conda packages.

To use 64-bit BLAS/LAPACK integers (ILP64 / BLAS64) with any supported BLAS/LAPACK library, install with:

python -m pip install -Csetup-args=-Duse_blas64=true .

If you combine BLAS64 with the MKL Pardiso backend, SCS requires 64-bit SCS integers as well (the default in the Meson build). At runtime SCS also checks that the process-wide MKL interface layer matches the LP64/ILP64 mode it was compiled for, and fails early if another library already set an incompatible MKL interface.

GPU

If you have a GPU and cuDSS installed you can install the GPU direct sparse solver using

python -m pip install -Csetup-args=-Dlink_cudss=true -Csetup-args=-Dint32=true .

See here for how to enable the GPU when solving. The sparse direct GPU solver is typically very fast.

See here for an example colab where the cuDSS version of SCS, along with required dependencies, is installed and used.

Spectral cones

To enable spectral cone support (log-determinant, nuclear norm, \(\ell_1\) norm, sum-of-largest-eigenvalues), install with:

python -m pip install -Csetup-args=-Duse_spectral_cones=true .

This requires LAPACK (enabled by default). See python_spectral_cone_keys for the cone dict keys.

Testing

To test that SCS installed correctly, and you have pytest installed, run

python -m pytest .

See here for the full SCS python API.

Legacy options

You can install with OpenMP parallelization support using

python legacy_setup.py install --scs --openmp

You can install the GPU indirect solver using

python legacy_setup.py install --scs --gpu