The course begins by providing a broad motivation for the design of multi- and many-core processors such as graphics processing units (GPUs). The first phase covers basic GPU concepts, including the evolution of GPU computing, a high-level overview of GPU architecture, and key differences from CPU architecture. Supplementary material is provided for students without a background in computer architecture fundamentals. Once a basic understanding is established, the module introduces the programming of GPUs using CUDA. The curriculum then delves into advanced concepts from both an architectural and programming perspective. To reinforce the theoretical knowledge gained in the lectures, students will engage in progressively more complex exercises throughout the module.
GPU Architectures and Programming
GPU Architectures and Programming
Description
The course begins by providing a broad motivation for the design of multi- and many-core processors such as graphics processing units (GPUs). The first phase covers basic GPU concepts, including the evolution of GPU computing, a high-level overview of GPU architecture, and key differences from CPU architecture. Supplementary material is provided for students without a background in computer architecture fundamentals. Once a basic understanding is established, the module introduces the programming of GPUs using CUDA. The curriculum then delves into advanced concepts from both an architectural and programming perspective. To reinforce the theoretical knowledge gained in the lectures, students will engage in progressively more complex exercises throughout the module.
- Study format
- Hybrid
- Application period
- 23 February – 22 March 2024
- Study period
- 2 April – 10 July 2024
- Credits
- 3 ECTS
- Pace
- 20%
- Hosting university
- Hamburg University of Technology
Learning outcomes
Understanding of the architecture of massively parallel GPUs
At the end of the course learner will gained a comprehensive understanding of the underlying architecture of massively parallel Graphics Processing Units (GPUs).
ESCO SKILLS
Effective GPU programming
By the end of the course lerners will be equiped with the skills necessary for effective GPU programming and be able to do programming.
ESCO SKILLS
Solve computationally intensive tasks faster and with lower energy consumption
By the end of the course students will be able to solve computationally intensive tasks faster and with lower energy consumption.
ESCO SKILLS
Potential progress
This graph shows the potential you could reach if you take this learning opportunity and how it fits your goal. You can also make a comparison to your current and potential competencies by pressing the buttons.
Entrepreneurship, technology and innovation
We recommend to turn your device to view graph
Energy and sustainability
Not sure which competencies suit you?
Take our motivation scan to find learning opportunities that will help you reach your potential goal and growth.
Information
The course begins by providing a broad motivation for the design of multi- and many-core processors such as graphics processing units (GPUs). The first phase covers basic GPU concepts, including the evolution of GPU computing, a high-level overview of GPU architecture, and key differences from CPU architecture. Supplementary material is provided for students without a background in computer architecture fundamentals. Once a basic understanding is established, the module introduces the programming of GPUs using CUDA. The curriculum then delves into advanced concepts from both an architectural and programming perspective. To reinforce the theoretical knowledge gained in the lectures, students will engage in progressively more complex exercises throughout the module.
Hosting university
Hamburg University of Technology
