blis
The Blis BLAS-like linear algebra library, as a self-contained C-extension.
Decision gist · record as of 2026-08-14
Yes, if you need fast BLAS-like operations in Python or Cython for inference or numerical workloads. The package is actively maintained, carries no known vulnerabilities, and offers both high-level and low-level APIs. Install friction is moderate due to compilation, but pre-built wheels cover most common platforms; non-standard architectures require manual BLIS_ARCH configuration.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- For non-standard CPU architectures, source installation requires LLVM on Windows and manual BLIS_ARCH environment variable configuration; pre-built wheels cover x86_64 and common ARM variants.
- Medium install friction due to compiled C extension requiring platform-specific wheels; wheels are pre-built for common architectures (x86_64, ARM, aarch64 on Linux, macOS, Windows), but source builds on non-standard architectures require LLVM on Windows and manual BLIS_ARCH configuration.
- Repository is actively maintained with recent commits.
License · maintenance · safety
BSD (permissive) — BSD license is permissive, allowing use in commercial and proprietary projects with minimal restrictions beyond attribution.
last release 2025-11-17 (270 days) · last repo commit 2026-05-14 · 238 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 21,857,473 downloads/mo, #989 on PyPI
Alternatives
Verify before relying
pip install blis
import blis
# High-level Python API available; Cython API requires cimport blis.cy- Whether the high-level Python API is documented with concrete function signatures and examples beyond the Cython interface shown in the description.
- Performance benchmarks or typical speedup factors compared to NumPy or other BLAS libraries for common operations.
What it is and what it does
blis is a Python C extension that exposes the Blis linear algebra library for fast matrix and vector operations. It wraps the underlying Blis routines (used in production ML inference) and offers two interfaces: a high-level Python API for general use, and direct Cython bindings with fused types and nogil support for performance-critical code. The package depends only on numpy and is optimized for single-threaded workloads, making it suitable for inference pipelines where thread overhead would be counterproductive.
Installation is straightforward on common platforms via pre-built wheels, but the package requires manual configuration for non-standard CPU architectures via the BLIS_ARCH environment variable. The repository is actively maintained, supporting Python 3.9 through 3.14, and carries no known security vulnerabilities.
Use it for
- Accelerate matrix multiplication (gemm) and other BLAS operations in ML inference pipelines where single-threaded execution is preferred.
- Write high-performance Cython code that calls Blis routines with fused types and nogil semantics for tight numerical loops.
- Replace generic NumPy linear algebra with optimized Blis implementations when inference latency is critical.
- Build custom numerical libraries that depend on fast, self-contained BLAS-like operations without external system BLAS dependencies.
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes, if you need fast BLAS-like operations in Python or Cython for inference or numerical workloads.
The package is actively maintained, carries no known vulnerabilities, and offers both high-level and low-level APIs. Install friction is moderate due to compilation, but pre-built wheels cover most common platforms; non-standard architectures require manual BLIS_ARCH configuration.
Install
blis on PyPI
Before you install
Medium install friction due to compiled C extension requiring platform-specific wheels; wheels are pre-built for common architectures (x86_64, ARM, aarch64 on Linux, macOS, Windows), but source builds on non-standard architectures require LLVM on Windows and manual BLIS_ARCH configuration. Repository is actively maintained with recent commits.
For non-standard CPU architectures, source installation requires LLVM on Windows and manual BLIS_ARCH environment variable configuration; pre-built wheels cover x86_64 and common ARM variants.
License in practice
BSD license is permissive, allowing use in commercial and proprietary projects with minimal restrictions beyond attribution.
Quickstart
pip install blis
import blis
# High-level Python API available; Cython API requires cimport blis.cyVerify before relying
- Whether the high-level Python API is documented with concrete function signatures and examples beyond the Cython interface shown in the description.
- Performance benchmarks or typical speedup factors compared to NumPy or other BLAS libraries for common operations.
Package facts
| License | BSD permissive |
| Python support | Supports the current Python release <3.15,>=3.9 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 1 packagenumpy |
| Maintenance | Actively maintained 270 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 21,857,473 / month, #989 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 4 - BetaEnvironment :: ConsoleIntended Audience :: DevelopersIntended Audience :: Information TechnologyLicense :: OSI Approved :: BSD LicenseOperating System :: MacOS :: MacOS XOperating System :: POSIX :: LinuxProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14Topic :: Scientific/Engineering |
Evidence: blis-1.3.3-cp310-cp310-macosx_10_9_x86_64.whl; blis-1.3.3-cp310-cp310-macosx_11_0_arm64.whl; blis-1.3.3-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; blis-1.3.3-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; blis-1.3.3-cp310-cp310-musllinux_1_2_aarch64.whl; blis-1.3.3-cp310-cp310-musllinux_1_2_x86_64.whl; blis-1.3.3-cp310-cp310-win_amd64.whl; blis-1.3.3-cp311-cp311-macosx_10_9_x86_64.whl; blis-1.3.3-cp311-cp311-macosx_11_0_arm64.whl; blis-1.3.3-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; blis-1.3.3-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; blis-1.3.3-cp311-cp311-musllinux_1_2_aarch64.whl; blis-1.3.3-cp311-cp311-musllinux_1_2_x86_64.whl; blis-1.3.3-cp311-cp311-win_amd64.whl; blis-1.3.3-cp312-cp312-macosx_10_13_x86_64.whl; blis-1.3.3-cp312-cp312-macosx_11_0_arm64.whl; blis-1.3.3-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl; blis-1.3.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl; blis-1.3.3-cp312-cp312-musllinux_1_2_aarch64.whl; blis-1.3.3-cp312-cp312-musllinux_1_2_x86_64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “blas linear algebra python”
- blisProvides fast BLAS-like linear algebra operations from Python and…
- onemkl-sycl-lapackProvides optimized C and Data Parallel C++ interfaces to Intel's…
- scipy-openblas32Provides pre-built OpenBLAS shared libraries and headers for use as a…
Give your agent the search over MCP, or paste the wish link into any chat.
More Scientific/Engineering packages
NumPy provides an N-dimensional array object and a comprehensive suite of mathematical, linear algebra, Fourier transform, and random number functions for scientific computing in Python.
pandas provides fast, flexible data structures (Series and DataFrame) for loading, cleaning, transforming, and analyzing labeled or relational data in Python.
scipy provides numerical algorithms for mathematics, science, and engineering—including optimization, integration, linear algebra, Fourier transforms, signal and image processing, and ODE solvers—built on numpy arrays.
scikit-learn provides a comprehensive Python library for supervised and unsupervised machine learning, including classification, regression, clustering, dimensionality reduction, and model evaluation tools built on NumPy and SciPy.
Install it if you need to train, evaluate, or deploy supervised or unsupervised learning models.
dill extends Python's pickle module to serialize and deserialize a much wider range of Python objects, including functions, lambdas, classes, and interpreter sessions, to byte streams for storage or network transmission.
Multiprocess is an enhanced fork of Python's standard multiprocessing library that uses dill for better serialization, allowing you to spawn processes with a threading-like API and share complex objects between them.
Install it if you use multiprocessing and encounter pickle serialization limits with lambdas or complex objects.
See also onemkl-sycl-blas · linear-operator · nvidia-cublas · nvidia-cublas-cu12 · onemkl-sycl-lapack · onemkl-sycl-sparse · cymem · scipy-openblas32 · nvidia-cusolver · nvidia-cusolver-cu12