Supercharge Python* applications and speed up core computational packages with this performance-oriented distribution

Python has become a pervasive and useful tool in advancing scientific research and computation. The language has very rich ecosystem of open source packages for mathematics, science, and engineering, anchored on the performant numerical computation on arrays and matrices, data analysis and visualization capabilities, and an interactive development environment that enables rapid and collaborative iteration of ideas.

The Intel® Distribution for Python is a ready-to-use, integrated package that delivers faster application performance on Intel® platforms. Performance accelerated packages with Intel Python enable scientists to take advantage of the productivity of Python, while taking advantage of the ever-increasing performance of modern hardware. Intel’s optimized implementations of NumPy and SciPy leverage the Intel® Math Kernel Library to achieve highly efficient multi-threading, vectorization, and memory management. Intel’s optimized implementation of mpi4py leverages the Intel® MPI Library to scale scientific computation efficiently across a cluster.

What’s New in the 2020 Release

The new release offers many performance improvements, including:

• Updated Python version to 3.7
• Faster machine learning with scikit-learn key algorithms accelerated with Intel® Data Analytics Acceleration Library
• Daal4py package improvements help address the needs of data scientists to harness Intel DAAL capabilities with a Pythonic API

Intel Optimized Packages for Intel® Distribution for Python

The Intel® Distribution for Python ships with many specialized packages that offer accelerated workflows and advanced functionality. A few of these packages are listed below:

Numerical and Scientific

• NumPy - The most popular numerical library for Python, accelerated with the Intel® MKL
• SciPy - The de-facto standard for a scientific toolset in the Python language, accelerated with the Intel® MKL
• numba - A Just-In-Time Compiler for decorated Python code that allows latest SIMD features and multi-core execution in order to fully utilize modern CPUs
• numexpr - A Python interface to symbolic and algebraic acceleration, via the Intel® MKL

Data Analytics

• Scikit-learn - A popular machine learning Python package, now accelerated with Intel's highest performance libraries.
• Pre-built and accelerated with Intel® MKL, Intel® DAAL, and Intel® Thread Building Blocks through direct source code changes to the package
• pyDAAL - A package for Python bindings to the Intel® Data Analytics Acceleration Library.
• Delivers a Python-interfaced solution for many of the steps in a data analytics pipeline, such as pre-processing, data transformations, dimensionality reduction, data modeling, prediction, and several drivers for reading and writing in most of the common data formats.
• Supports many computation modes, including Batch, Distributed, and Online modes for many of the support algorithms.

Parallelism

• smp - Static Multi-Processing, a module handling nested parallelism issues like oversubscription while composing different parallel components
• tbb - A module controlling and enabling the use of Intel® Thread Building Blocks for dynamic orchestration of different parallel components
• mpi4py - A Python interface to MPI, through Intel® MP

For a complete list of packages, please visit the Complete List of Packages for the Intel® Distribution for Python*

Intel® Optimizations improve Python scikit-learn efficiency closer to native code speeds on Intel Xeon® processors