Want to keep learning?

This content is taken from the Partnership for Advanced Computing in Europe (PRACE)'s online course, Python in High Performance Computing. Join the course to learn more.

Hands-on: Optimising heat equation solver

In this exercise you can practice different optimization techniques with the heat equation solver.

The code for the exercise is located under under cython/heat-equation.

Creating a Cython extension

Write a setup.py for creating a Cython version of heat.py module, and use it from the main program heat_main.py. How much does simple Cythonization (i.e. diminishing the interpreting overhead) improve the performance?


Based on the profile in the performance measurement exercise optimise the most time consuming part of the algorithm. If you did not finish the profiling exercise, you can look at example profile here.

Utilize all the tricks you have learned so far (type declarations, fast array indexing, compiler directives, C functions, …).

Investigate and comment how the different optimizations affect the performance. You can use applications own timers and/or timeit. Annotated HTML-report with cython -a … can be useful when tuning performance.

When finished with the optimisation, compare the performance to Python/NumPy model solution (in numpy/heat-equation), which uses array operations. You can play around also with larger input data as provided in bottle_medium.dat and bottle_large.dat.

Share this article:

This article is from the free online course:

Python in High Performance Computing

Partnership for Advanced Computing in Europe (PRACE)