Intel Distribution for Python is available for all modorn operating systems including Windows, Linux, and MacOS. For me, it is worth to try this Python distribution for your projects, especially scientific computing projects. By the way, it is free of charge.

Here are the optimized Python packages developed by Intel on top its original version.

• Numerical and Scientific: NumPy, SciPy, Numba, numexpr
• Data Analytics: scikit-learn, pyDAAL, daal4py
• Parallelism: smp, tbb, mpi4py

This software is mainly developed for

• Machine Learning Developers, Data Scientists, and Analysts
• Numerical and Scientific Computing Developers
• High-Performance Computing (HPC) Developers

As the installation of Intel Distribution for Python has been well described at its homepage, select the OSs on which you want Python to be installed. The installer can be download from https://software.seek.intel.com/python-distribution.

If using the stand-alone installer, complete the following steps:

1. Change the directory to the installation path.
2. Download and then unzip the the zip file.
3. Change directory to intelpython2 or intelpython3 (depending on the version you’ve downloaded).
4. Double-click setup_intel_python.bat or execute that script from the command prompt.
5. When the installation completes, activate your root Intel python conda environment.

To modify only your current command shell, use the .\Scripts\activate command.

If using the stand-alone installer, complete the following steps:

1. Change directory to the installation path.
2. Download the tarball for the Intel Distribution for Python
3. Extract the contents using the tar -xvzf command
4. Change directory to intelpython2 or intelpython3 (depending on the version you’ve downloaded)
5. Run from shell: bash setup_intel_python.sh
6. When the installation completes, activate your root Intel python conda environment:
   • To modify only your current shell, use the source ./bin/activate command
   • To modify all future logins, do one of the following:
     • Add source <install>/bin/activate root to your .bashrc (bash) or other logon script
     • Manually add the <install>/bin directory to your PATH
   • To ensure your environment points to Intel Distribution for Python, run the which pythoncommand.

1. Change directory to the installation path
2. Download the tarball for the Intel Distribution for Python
3. Extract the contents using the tar -xvzf command.
4. Change directory to intelpython2 or intelpython3 (depending on the version you’ve downloaded)
5. Run from shell: bash setup_intel_python.sh
6. When the installation completes, activate your root Intel python conda environment:
   • To modify only your current shell, use the source ./bin/activate command
   • To modify all future logins, do one of the following:
     • Add source <install>/bin/activate root to your .bashrc (bash) or other logon script
     • Manually add the <install>/bin directory to your PATH
   • To ensure your environment points to Intel® Distribution for Python*, run the which python command.

You can use both PIP or Conda to install full pacakages or individial packages. However, I found that the Intel's packages available on Conda server is a bit updated ahead of those on PIP. So if you are using Anaconda's Python distribution, I strongly suggest you use conda.

To avoid confusion between projects and packages, we can command conda to create us a new environment, install specific version of Python, and packages at the same time. The following steps are example of installation of Intel's Python and some packages.

1. Update conda

conda update conda

2. Create new empty environment, Python 3, and Intel's Python 3.

conda create -n intel -c intel intelpython3_full python=3

3. Enter environment

activate intel

4. Install packages

conda install numpy scipy scikit-learn pydaal tbb4py

Rangsiman Ketkaew