Prerequisites
Should have experience programming in at least one language, ideally Python. If you do not have any experience programming in Python it is advisable to have a look at the Software Carpentry lesson on Python which introduces the basics of Python.
| Setup | Setup required for the lesson | ||
| Day 1 | 14:00 | 1. Introduction |
What is Dask?
How does it work? |
| 14:15 | 2. Python |
What is a virtual environment?
How can you time Python scripts? What is a simple way to run a Python script on a compute node? |
|
| 14:40 | 3. Break | Break | |
| 14:55 | 4. Dask Delayed |
How do you install Dask Delayed?
How can Dask Delayed be used to parallelize python code? How can I see what Dask Delayed is doing? |
|
| 15:35 | 5. Real Computations |
What happens when I use Dask for “real” computations?
How can I see CPU efficiency of a SLURM Job? |
|
| 16:00 | Finish | ||
| Day 2 | 14:00 | 6. Distributed Computations |
How can I avoid the GIL problem?
How can I run multiple Python interpreters at once on one problem? |
| 14:30 | 7. Cython |
How slow is Python?
Can I make my Python code faster? How can I turn my Python code into compiled C or C++ code? |
|
| 15:10 | 8. Break | Break | |
| 15:25 | 9. NumPy |
Why are NumPy arrays faster than lists?
How do you create NumPy arrays? What are some things I can do with NumPy arrays? |
|
| 15:40 | 10. Dask Array |
How can I process NumPy arrays in parallel?
Can I distribute NumPy arrays to avoid needing all the memory on one machine? |
|
| 16:05 | 11. Wrapping Up |
What did we learn?
What next? |
|
| 16:10 | Finish |
The actual schedule may vary slightly depending on the topics and exercises chosen by the instructor.