In this course, you will learn how to implement some widely-used algorithms as fast and scalable programs on modern hardware, and how to evaluate your implementation using appropriate test cases and performance experiments.
This course introduces the students to some classical examples of widely used algorithms and uses these to show how to implement an algorithm in a correct and scalable way in the context of a particular application.
The focus of the course is the implementation, its correctness, the experimental analysis, and the connection to the performance models.
Examples of well understood computational tasks that can be considered in depth are:
- shortest path for cars on road networks
- sorting: why different variants of quicksort differ in performance, how to sort strings
- cache oblivious search trees as file system or database index
- matrix multiplication
- matrix vector multiplication
- locality sensitive hashing
- k-means clustering
- minimum spanning tree
- edit distance
- approximate neighborhood
- distinct elements
- job scheduling
The course introduces the different algorithmic tasks, the algorithms, the testing methodology and the performance models. The exercises focus on implementation, testing and interpretation of the measurements
Intended learning outcomes
After the course, the student should be able to:
- implement a (randomized, parallel) algorithm from a given high-level description (such as pseudo-code). In particular, this includes judging which data structures, libraries, frameworks, programming languages, and hardware platforms are appropriate for the computational task, and using them effectively in the implementation.
- experimentally evaluate the correctness of an implementation by creating and using adequate test cases.
- design, perform, and run a performance and scalability experiment. In particular, this includes creating and using adequate families of synthesized or real-world test cases, measuring the computational resource requirements of a particular implementation, and judging if the experimental results conform to a theoretical performance model.
Ordinary examExam type:
D: Submission of written work with following oral, external (7-trinsskala)
D2G: Submission of written work for groups with following oral exam supplemented by the work submitted. The group has a shared responsibility for the content of the report.
Mixed 2 Exam:
The group makes their
and afterwards the
students participate in the
dialogue individually while
the rest of the group is
outside the room.
Hand-in: Implementation and experimental evaluation of a self-developed algorithm, or an algorithm as described in one or several research papers.
Duration of oral exam: 20 minutes per student
Group size: 1-3