Python in scientific computing: what works and what doesn't


Follow to receive video recommendations   a   A

There is no want of technologies for doing scientific calculations in Python. In this talk I will share some hard-learned knowledge about what works and what doesn’t with the libraries we are using at GEM (the Global Earthquake Model foundation). I will show how the following libraries fare with respect to our main concerns of performance, simplicity, reliability and portability

  • h5py
  • celery/rabbitmq
  • PyZMQ
  • numpy/scipy
  • rtree

and I will talk about several library bugs we found and had to work around. I will also talk about some libraries that we do not use (such as cython, numba, dask, pytables, …) and the reason why we do not use them. Hopefully this will be useful to people using or planning to use a similar software stack.

My slides are here:

Editors Note:

I would like to work with open source projects to create a branch of the tree with all of the best videos for your open source project. Please send me an email if you are interested.