The projects are all about parallelizing an algorithm or data structure from Programmering FK, EDAA01, or Algorithms, data structures and complexity, EDAF05:
- Parallel breadth-first search: Emil Bengtsson
- Parallel Quicksort: Felix Lundström
- Parallel hash tables: Lukas Forsberg & Edvin Andersson.
- Parallel computational fluid dynamics: Martin Karp
- Parallel version of your favorite other algorithm or data structure
The project will be done in groups of one or two students as follows:
- Read papers about a proposed algorithm
- Implement it in steps from sequential, basic parallel, to faster parallel
- While implementing your algorithms, reflect on what is complicated to make correct and efficient
- Evaluate it in terms of load balancing and scalability e.g. using from 1 to 80 threads on POWER8. If you prefer to use your own computer with a GPU then that is also fine.
- Investigate to what extent transactional memory or atomic variable are suitable to use
- Write a report with your findings
- Make a presentation
Presentations on December 19, at 9-12 in E:2405.
- Emil Bengtsson: Parallel breadth-first search
- Lukas Forsberg and Edvin Andersson: Parallel hash table
- Martin Jacobsson: Parallel sorting: Swapsort
- Martin Karp: Parallel computational fluid dynamics
- Felix Lundström: Parallel Quicksort
- Daniel Regefalk and Pontus Olsson: Indoor localisation based on WiFi signal signatures and Deep Learning
- Erik Samuelsson: title to be decided
- Jonathan Strandberg: title to be decided