These are the materials for an introductory workshop on Spark, Chapel, and TensorFlow - technologies of immediate interest to researchers performing large-scale numerical computing, who may be using MPI currently.
The presentation can be viewed at http://ljdursi.github.io/Spark-Chapel-TF-UMich-2016
The intent is that attendees with some experience in scientific computation and data analysis and some familiarity with python will, after working through this material:
- Know how to perform some basic canonical scientific computations in each model
- Understand under what regimes MPI, Spark, Chapel, and TensorFlow are best
Included in the materials are the files for a Vagrant VM which contains:
- Spark, and the Jupyter Notebook for Spark
- The Chapel compiler
- The TensorFlow runtime
- Examples in each model