Introduction to GPU and accelerator programming for scientific computing
Course main content
Graphics processing units (GPU) and other accelerators will provide a significant part of the computational power used in future scientific computing applications. Mastering parallel computing with these devices is therefore very interesting and becoming increasingly important. This course will provide knowledge and practical experience for the development of programs using GPUs and accelerators in the field of scientific computing.
The lectures focus on introducing programming with CUDA on GPUs as well as programming approaches for the Intel MIC architecture. More examples of alternative processor architectures like the Parallella platform and multi-core digital signal processors will be discussed in order to round the picture off. Practical exercises during the second course week will exemplify the features of the computing devices. Finally, one numerical simulation application will be developed during the third week.
Students will after the course be able
- To understand the properties of GPUs and other accelerator devices
- To reason about the performance of programs running on it
- To assess the potential and limitations of using this computing approach
- To write parallel programs for GPU and Xeon Phi processors
Participation in lectures and presence labs
Week 6/2014 Lecture week (presence at KTH)
Week 7/2014 Self-studies and assignments
Week 8/2014 Project work
Michael Schliephake, michs at pdc.kth.se
Send an e-mail to michs at pdc.kth.se with the following information:
- SeSE, “Introduction to GPU and Accelerator programming”
- e-mail (You must use your university email address, not gmail, yahoo, hotmail etc.)
- Subject of PhD-project.
Indicate also if you apply for a travel grant.
Deadline for application: 20 January 2014