Search
News
Youtube Channel CG
You are here
GPU Programming (previously: Computer Graphics IV)
Semester:
2021-2022 Wintersemester
Lecturer:
Tutor:
Location/Time:
Mo 9-10 H-C 7326
SWS/CP:
1+2 / 6 (oder 5)
Recommended for:
Master-Inf
Examination Type:
Oral 20-40 minutes; prerequisite: project task
Announcements:
Please register in Unisono for this course. You will then get a mail at the start of the term containing information about enrolling in the Moodle Course for this lecture.
If you have any questions, please contact Martin Lambers.
Goals:
This lecture is about programming massively parallel processors, in particular graphics processing units (GPUs), using generic APIs and methods that are independent from computer graphics. Students do not need computer graphics knowledge, but should have first experiences with C/C++ programming.
This is mainly a practical course, with one hour of lecture and two hours of hands-on exercises per week, plus a larger project task in the second half of the term.
Contents:
- General Purpose Computations on Graphics Processing Units (GPGPU)
- Graphics-independent APIs for programming GPUs: mainly OpenCL and CUDA, plus an overview of OpenACC and SYCL
- Selected algorithms for massively parallel processors
Literature:
- Kirk, Hwu: Programming Massively Parallel Processors: A Hands-on Approach, Third Edition, Morgan Kaufman 2017
Tutor:
Tutorial Location / Tutorial Time:
Mo 10-12 H-A 7118
Tutorial:
The lab exercise illustrates and deepens the material by means of programming examples.
An additional cross-section exercise allows the combination of several techniques in one's own project.
Script, example code, exercises and other material will be available from the Moodle site.