Mike McKerns – Efficient Python for High-Performance Parallel Computing – PyCon 2016

Speaker: Mike McKerns

This tutorial is targeted at the intermediate-to-advanced Python user who wants to extend Python into High-Performance Computing. The tutorial will provide hands-on examples and essential performance tips every developer should know for writing effective parallel Python. The result will be a clear sense of possibilities and best practices using Python in HPC environments.

Slides can be found at: and

Similar Posts

3 thoughts on “Mike McKerns – Efficient Python for High-Performance Parallel Computing – PyCon 2016
  1. I recall various mentions of both avoiding for loops and that for loops are no longer the performance killer that they once were. Both are mentioned in this presentation, which moved finding those references higher in my devastatingly long TODO list.

    Wave there been recognizable moments when for loops became radically cheaper? Or has been a gradual journey? Looking through release notes hasn't done it for me: the closest I can come is PEP 234 — Iterators. That's Python 2.1, from 2001-01-30!

  2. The presentation ends with the displayed text "The other end of the spectrum is high-performance parallel instead of large-scale parallel". What is meant by this? MPI vs multiprocessing, perhaps?

  3. "n in set(xrange(x))" and "n in list(xrange(x))" are both suboptimal because "n in xrange(x)" works and is expected to be O(1). And in Python 3, of course that's "n in range(x)".

Leave a Reply

Your email address will not be published. Required fields are marked *