mkl-static provides the static libraries for Intel® oneAPI Math Kernel Library (MKL). MKL is a highly optimized library of mathematical functions (BLAS, LAPACK, FFTs, etc.) essential for high-performance computing. It is primarily used as a backend for scientific Python libraries like NumPy and SciPy to deliver maximum performance, rather than being directly imported in Python. The current version is 2025.3.1, and it's maintained by Intel as part of their oneAPI ecosystem.
pip install mkl-staticNo compatibility data collected yet for this library.
mkl-static does not provide direct Python imports. Its quickstart involves demonstrating how a Python library, such as NumPy, is configured to use the MKL backend. This code snippet prints NumPy's configuration, which should show 'mkl_info' if it's correctly linked to MKL.
Do not attempt to `import mkl-static`. Instead, ensure your numerical Python libraries are built or configured to use MKL. Check `numpy.__config__.show()` to verify MKL linkage.
Ensure that if you intend to use MKL, your Python packages (e.g., NumPy, SciPy) are installed from sources that are built against MKL (e.g., conda-forge often provides MKL-enabled builds) or explicitly configured to link to the `mkl-static` binaries. The `mkl-service` package can sometimes help manage MKL threading.
Use a dedicated virtual environment. If using Conda, let Conda manage MKL dependencies via `conda install numpy scipy mkl`. If using pip, ensure a consistent environment. Intel's `mkl-service` package provides utilities to manage MKL runtime behavior.
Ensure `mkl-static` is installed. Check that relevant environment variables like `LD_LIBRARY_PATH` (Linux) or `DYLD_LIBRARY_PATH` (macOS) include the MKL library path, or use `conda install mkl-static` within a conda environment, which typically handles library paths automatically. Alternatively, explicitly link the dependent library against MKL.
Run `import numpy as np; np.__config__.show()`. Look for `mkl_info` to confirm MKL linkage. If absent, you might need to reinstall NumPy/SciPy from a source that explicitly builds against MKL (e.g., `conda install numpy scipy mkl` or using specific wheels that bundle MKL).
Ensure that `mkl-static` (or another MKL provider) is correctly installed and its libraries are discoverable in your system's PATH (Windows) or LD_LIBRARY_PATH (Linux). If using Conda, `conda install mkl-service` usually resolves this by pulling in necessary MKL dependencies.
No dependency data recorded yet.