Building from source#

Note

If you are only trying to install SciPy, we recommend using binaries - see Installation for details on that.

Building SciPy from source requires setting up system-level dependencies (compilers, BLAS/LAPACK libraries, etc.) first, and then invoking a build. The build may be done in order to install SciPy for local usage, develop SciPy itself, or build redistributable binary packages. And it may be desired to customize aspects of how the build is done. This guide will cover all these aspects. In addition, it provides background information on how the SciPy build works, and links to up-to-date guides for generic Python build & packaging documentation that is relevant.

System-level dependencies#

SciPy uses compiled code for speed, which means you need compilers and some other system-level (i.e, non-Python / non-PyPI) dependencies to build it on your system.

Note

If you are using Conda, you can skip the steps in this section - with the exception of installing compilers for Windows or the Apple Developer Tools for macOS. All other dependencies will be installed automatically by the mamba env create -f environment.yml command.

If you want to use the system Python and pip, you will need:

  • C, C++, and Fortran compilers (typically gcc, g++, and gfortran).

  • Python header files (typically a package named python3-dev or python3-devel)

  • BLAS and LAPACK libraries. OpenBLAS is the SciPy default; other variants include ATLAS and MKL.

  • pkg-config for dependency detection.

To install SciPy build requirements, you can do:

sudo apt install -y gcc g++ gfortran libopenblas-dev liblapack-dev pkg-config python3-pip python3-dev

Alternatively, you can do:

sudo apt build-dep scipy

This command installs whatever is needed to build SciPy, with the advantage that new dependencies or updates to required versions are handled by the package managers.

To install SciPy build requirements, you can do:

sudo dnf install gcc-gfortran python3-devel openblas-devel lapack-devel pkgconfig

Alternatively, you can do:

sudo dnf builddep scipy

This command installs whatever is needed to build SciPy, with the advantage that new dependencies or updates to required versions are handled by the package managers.

To install SciPy build requirements, you can do:

sudo yum install gcc-gfortran python3-devel openblas-devel lapack-devel pkgconfig

Alternatively, you can do:

sudo yum-builddep scipy

This command installs whatever is needed to build SciPy, with the advantage that new dependencies or updates to required versions are handled by the package managers.

To install SciPy build requirements, you can do:

sudo pacman -S gcc-fortran openblas pkgconf

Install Apple Developer Tools. An easy way to do this is to open a terminal window, enter the command:

xcode-select --install

and follow the prompts. Apple Developer Tools includes Git, the Clang C/C++ compilers, and other development utilities that may be required.

Do not use the macOS system Python. Instead, install Python with the python.org installer or with a package manager like Homebrew, MacPorts or Fink.

The other system dependencies you need are a Fortran compiler, BLAS and LAPACK libraries, and pkg-config. They’re easiest to install with Homebrew:

brew install gfortran openblas pkg-config

Note

As of SciPy >=1.2.0, we do not support compiling against the system Accelerate library for BLAS and LAPACK. It does not support a sufficiently recent LAPACK interface. This is planned to change in 2023, because macOS 13.3 introduced a major upgrade to Accelerate which resolved all known issues.

A compatible set of C, C++ and Fortran compilers is needed to build SciPy. This is trickier on Windows than on other platforms, because MSVC does not support Fortran, and gfortran and MSVC can’t be used together. You will need one of these sets of compilers:

  1. Mingw-w64 compilers (gcc, g++, gfortran) - recommended, because it’s easiest to install and is what we use for SciPy’s own CI and binaries

  2. MSVC + Intel Fortran (ifort)

  3. Intel compilers (icc, ifort)

Compared to macOS and Linux, building SciPy on Windows is a little more difficult, due to the need to set up these compilers. It is not possible to just call a one-liner on the command prompt as you would on other platforms.

First, install Microsoft Visual Studio - the 2019 Community Edition or any newer version will work (see the Visual Studio download site). This is needed even if you use the MinGW-w64 or Intel compilers, in order to ensure you have the Windows Universal C Runtime (the other components of Visual Studio are not needed when using Mingw-w64, and can be deselected if desired, to save disk space).

There are several sources of binaries for MinGW-w64. We recommend the RTools versions, which can be installed with Chocolatey (see Chocolatey install instructions here):

choco install rtools -y --no-progress --force --version=4.0.0.20220206

In case of issues, we recommend using the exact same version as used in the SciPy GitHub Actions CI jobs for Windows.

The MSVC installer does not put the compilers on the system path, and the install location may change. To query the install location, MSVC comes with a vswhere.exe command-line utility. And to make the C/C++ compilers available inside the shell you are using, you need to run a .bat file for the correct bitness and architecture (e.g., for 64-bit Intel CPUs, use vcvars64.bat).

For detailed guidance, see Use the Microsoft C++ toolset from the command line.

Similar to MSVC, the Intel compilers are designed to be used with an activation script (Intel\oneAPI\setvars.bat) that you run in the shell you are using. This makes the compilers available on the path. For detailed guidance, see Get Started with the Intel® oneAPI HPC Toolkit for Windows.

Note

Compilers should be on the system path (i.e., the PATH environment variable should contain the directory in which the compiler executables can be found) in order to be found, with the exception of MSVC which will be found automatically if and only if there are no other compilers on the PATH. You can use any shell (e.g., Powershell, cmd or Git Bash) to invoke a build. To check that this is the case, try invoking a Fortran compiler in the shell you use (e.g., gfortran --version or ifort --version).

Building SciPy from source#

If you want to only install SciPy from source once and not do any development work, then the recommended way to build and install is to use pip:

If you are using a conda environment, pip is still the tool you use to invoke a from-source build of SciPy. It is important to always use the --no-build-isolation flag to the pip install command, to avoid building against a numpy wheel from PyPI. In order for that to work you must first install the remaining build dependencies into the conda environment:

# Either install all SciPy dev dependencies into a fresh conda environment
mamba env create -f environment.yml

# Or, install only the required build dependencies
mamba install python numpy cython pythran pybind11 compilers openblas pkg-config

# To build the latest stable release:
pip install scipy --no-build-isolation --no-binary scipy

# To build a development version, you need a local clone of the SciPy git repository:
git clone https://github.com/scipy/scipy.git
git submodule update --init
pip install . --no-build-isolation
# To build the latest stable release:
pip install scipy --no-binary scipy

# To build a development version, you need a local clone of the SciPy git repository:
git clone https://github.com/scipy/scipy.git
git submodule update --init
pip install .

Building from source for SciPy development#

If you want to build from source in order to work on SciPy itself, first clone the SciPy repository:

git clone https://github.com/scipy/scipy.git
git submodule update --init

Then you want to do the following:

  1. Create a dedicated development environment (virtual environment or conda environment),

  2. Install all needed dependencies (build, and also test, doc and optional dependencies),

  3. Build SciPy with our dev.py developer interface.

Step (3) is always the same, steps (1) and (2) are different between conda and virtual environments:

Note

There are many tools to manage virtual environments, like venv, virtualenv/virtualenvwrapper, ``pyenv/pyenv-virtualenv, Poetry, PDM, Hatch, and more. Here we use the basic venv tool that is part of the Python stdlib. You can use any other tool; all we need is an activated Python environment.

Create and activate a virtual environment in a new directory named venv ( note that the exact activation command may be different based on your OS and shell - see “How venvs work” in the venv docs):

python -m venv venv
source venv/bin/activate

Then install the Python-level dependencies (see pyproject.toml) from PyPI with:

# Build dependencies
python -m pip install numpy cython pythran pybind11 meson ninja pydevtool rich-click

# Test and optional runtime dependencies
python -m pip install pytest pytest-xdist pytest-timeout pooch threadpoolctl asv gmpy2 mpmath

# Doc build dependencies
python -m pip sphinx "pydata-sphinx-theme==0.9.0" sphinx-design matplotlib numpydoc jupytext myst-nb

# Dev dependencies (static typing and linting)
python -m pip mypy typing_extensions types-psutil pycodestyle ruff cython-lint

If you don’t have a conda installation yet, we recommend using Mambaforge; any conda flavor will work though.

To create a scipy-dev development environment with every required and optional dependency installed, run:

mamba env create -f environment.yml
mamba activate scipy-dev

Note

On Windows it is possible that the environment creation will not work due to an outdated Fortran compiler. If that happens, remove the compilers entry from environment.yml and try again. The Fortran compiler should be installed as described under the Windows tab of the System-level dependencies section higher up.

To build SciPy in an activated development environment, run:

python dev.py build

This will install SciPy inside the repository (by default in a build-install directory). You can then run tests (python dev.py test), drop into IPython (python dev.py ipython), or take other development steps like build the html documentation or running benchmarks. The dev.py interface is self-documenting, so please see python dev.py --help and python dev.py <subcommand> --help for detailed guidance.

Customizing builds#

Background information#