This is an umbrella topic related to our research on the efficient implementation of nested data parallelism. Students who are interested in parallel programs and their compilation, especially for PC clusters, can contact Gabriele Keller or Manuel Chakravarty for details.

Within the area of nested data parallelism, we do both practical and theoretical work. Practical work will involve the development of optimised libraries that support parallel applications as well as example applications as case studies. Theoretical will involve proving core properties of program transformations, possibly with the help of an automated theorem prover. Thesis topics can be either only practical, only theoretical, or both.

• Core requirements for successful completion:
  • For a practical topic, you need to be an experienced programmer.
  • For a theoretical topic, you should have a basic understanding of discrete math and be able to develop mathematical proofs.
  • Thesis-specific requirements will be developed on an individual basis.
• Extra tasks for the ambitious:
  • These will be developed on a case by case basis.

• A tutorial to efficient functional array programs.
• The corresponding research project.