Achieve Greater Performance with Minimal or No Code Changes

• Near-native performance comes through acceleration of core Python numerical packages. This is done using Intel® Performance Libraries such as Intel® oneAPI Math Kernel Library (oneMKL) and Intel® oneAPI Data Analytics Library (oneDAL).
• Compile your custom code using core compiler technologies such as Numba* and Cython. These apply vectorization and multithreading to your code, bringing Python to new levels.
• Leverage composable parallelism through Intel® oneAPI Threading Building Blocks (oneTBB) and symmetric multiprocessing (SMP). These technologies help you avoid excess headroom, use nested parallelism without oversubscription, and optimize particular workloads for the highest performance.
• Program for multiple devices using the same programming model, DPPy (Data Parallel Python), thereby avoiding rewriting CPU code to device code. Its included package, dpctrl, lets you control execution and data management across the host and devices.
• Run your numeric and machine learning (ML) workloads on data-parallel and heterogeneous hardware without having to go out of the Python ecosystem, while simultaneously not compromising on performance using DPPy.

Develop in the Cloud

Get what you need to build and optimize your oneAPI projects for free. With an Intel® DevCloud account, you get 120 days of access to the latest Intel® hardware—CPUs, GPUs, FPGAs—and Intel® oneAPI tools and frameworks. No software downloads. No configuration steps. No installations.

Get Access

Download the Library

Intel Distribution for Python is included as part of the Intel® AI Analytics Toolkit.

Get It Now

Installation Guides

Intel Installer

Documentation

• Empower Science with High-Performance Python*
• Faster and Scalable Machine Learning with Python for Data Scientists
• Optimized Packages for the Intel® Distribution for Python*
• Known Issues
• End User License Agreement (EULA)
• View All Documentation

Tutorials

• Unleash the Parallel Performance of Python Programs
• Get Started with Intel® Software Optimization for Theano*
• Composable Multiprocessing and Multithreading for Numeric Libraries
• Composable Multithreading for Python Libraries
• Faster Random Number Generation in Intel Distribution for Python

Videos

• Intel Distribution for Python: Highlights & Overview
• Introduction to Scalable DataFrame Compiler
• Speed Up Python Applications & Make Core Computations Soar
• Techniques to Accelerate NumPy & SciPy
• Benefits & Features of Profiling Python in Intel® VTune™ Profiler
• Take Advantage of Three Features in Persistent Memory Products
• Maximum Performance, Minimum Effort with Intel® Performance Libraries
• Accelerate Machine Learning
• Use Runtimes in the Data Center for Evolving Performance Benchmarks
• Accelerate scikit-learn* 
• Is Python Almost as Fast as Native Code? Believe It!
• Unlock Composable Parallelism in Python
• A Performance Analysis of Python Applications with Intel VTune Profiler 
• Boost Python Performance with Intel® oneAPI Math Kernel Library
• Remove Python Performance Barriers for Machine Learning

Operating systems:

• Linux*
• Windows® 10
• macOS*

Language:

• Python 3.7.4

Package management:

• conda*
• PIP*

Compatible with:

• Microsoft Visual Studio*
• PyCharm*

For more information, see the system requirements.