Graphics Processing Units (GPUs) are specialized hardware designed to accelerate the processing of graphics and visualizations. GPUs have become increasingly popular for a variety of non-graphics related tasks, including scientific computing, machine learning, and data analysis. Today, GPUs are also used for CFD (Computational Fluid Dynamics) and FEM (Finite Element Method). The high par-allelization capabilities of GPUs make them well-suited for CFD and FEM.
General information
The students will learn how to write a simple CFD or FEM code or a Poisson solver. The code should run entirely or partly on the GPU. PhD students are welcome.
Course content
Introduction lectures on CUDA programming including two mini-workshops.
Project 1
- The student groups write a simple CFD/FEM code or Poisson solver in CUDA. Ideally, each group includes students with knowledge in CFD, FEM or Poisson equation and CUDA
- Profiling (GPU time, uploading/downloading data to/from the GPU etc)
- Written and oral presentation of the project
More information about the course
Prerequisites
The students should have good knowledge in Python.
How to apply
Apply to all Tracks courses at universityadmissions.se/antagning.se.
At universityadmissions.se/antagning.se: Search for the course you are interested in by using the course code starting with TRA.
Read more here.
Please include a letter explaining your contribution to the project group. This may be used when prioritizing if we get too many applicants.
For alumni, PhD-stduents and professionals the course selection follows a different process. See more information on Tracks web page.
Details
Teachers:
Lars Davidson, Rickard Bensow, Fredrik Larsson, Miquel Pericas
Course dates: Study period 2, 2025
Credits: 7.5
Level: Advanced
Course code: TRA220
Application deadline: