Intel python and intel environment in anaconda difference?

Question

I am using Windows and I followed these steps to get the intel environment for my anaconda 3.

I realized that Intel Python can be obtained without anaconda by directly downloading it from the Intel website. (Intel python Download option screenshot)

So What is the difference in Intel python in both the cases?

Answer 1

There is no difference. You have the option of choosing a full installer to get the entire Intel python installation, or creating a conda environment to install just the packages you want. The article says to install intelpython<2|3>_core or intelpython<2|3_full. These are collections of packages, but you can also create an environment with just Intel's optimized numpy, for instance, like this:

conda create -n intel_np -c intel numpy

Answer 2

Anaconda is the leading Open Data Science platform powered by Python. Anaconda includes an open source core. Python distribution and over 700 packages. The Anaconda installer includes 150+ packages, and the balance of the packages are easily installed via the conda install package name command.

The Intel Distribution for Python Powered by Anaconda is focused on tools and techniques to accelerate the performance of Python and the widely used numerical computing packages, especially for the latest Intel processors. Both distributions work on Intel KNL, the recently released Intel Xeon Phi. To ensure compatibility, the Intel Distribution for Python Powered by Anaconda shares build recipes with Anaconda and uses conda for packaging. The Intel packages for accelerating Python performance are available on AnacondaCloud.

The Anaconda Distribution includes many more packages to provide a more complete ecosystem and focuses on supporting a wide user base that includes the latest as well as older hardware and operating systems. The Anaconda Distribution installers include over 150 packages and give users access to over 400 Python packages and over 200 R packages hosted on repo.continuum.io and installable with conda. The full list of packages available packages for the Anaconda Distribution is available on the Continuum website.

The Intel Distribution accelerates performance of certain numerical data science Python packages such as Numpy, Scipy, Scikit-learn, Tensorflow for cpu etc., with Intel® Performance Libraries, including Intel® Math Kernel Library (Intel® MKL), Intel® Threading Building Blocks (Intel® TBB), Intel® Data Analytics Acceleration Library (Intel® DAAL), and Intel® MPI. The packages have been optimized to take advantage of parallelism through the use of threading, multiple nodes, and vectorization.