Introduction to GPU Programming¶

Overview of Modular Lesson Material¶

Graphical Processing Units (GPUs) have become the primary computational workhorse of many modern high-performance computing (HPC) systems around the world. Over the past decade, the number of GPU-enabled supercomputers in the Top500 list of the world’s most powerful supercomputers has grown steadily, and this trend is expected to continue. As HPC architectures increasingly rely on GPUs and other hardware accelerators to deliver performance and energy efficiency, a large fraction of the computational resources available to researchers, engineers, and scientists will be provided by these devices. Consequently, the ability to develop efficient GPU-accelerated applications has become an essential skill for HPC software developers.

At the same time, the GPU hardware and software ecosystem has become increasingly diverse and complex, with a wide range of hardware platforms, software stacks, and programming environments available to software developers. Multiple hardware vendors compete in the accelerator market, each offering its own architectures, software stacks, programming models, and development toolkits. Beyond vendor-specific solutions, software developers can choose from a growing range of programming models, domain-specific frameworks, libraries, and compiler technologies. As a result, it can be challenging for software developers and project owners to navigate this complex hardware and software landscape and select the most appropriate GPU programming model or framework that best meets the needs of their projects. The choice depends on a range of factors, including application requirements, target hardware platforms, performance objectives, code portability, long-term maintainability, and compatibility with existing code.

This introductory module provides an overview of the contemporary GPU programming landscape. Rather than focusing on a single programming model, it introduces the major GPU hardware platforms, programming languages, frameworks, and portability solutions currently used in HPC. Participants will gain an understanding of the strengths, limitations, and typical use cases of different approaches, enabling them to make informed decisions when selecting technologies for new software projects or when adapting existing HPC applications for accelerator-based systems.

Prerequisites

A basic understanding of programming in C/C++ or Fortran.
Familiarity with compiling and running programs from the command line in a Linux environment.
A basic understanding of parallel computing concepts, such as processes, threads, and shared versus distributed memory, is beneficial but not required.
Some experience using HPC systems (e.g., logging in via SSH, compiling code, and submitting jobs to a scheduler) is helpful but not essential.
No prior experience with GPU programming is assumed.

Table of Content¶

Episodes

Reference

Learning Outcomes¶

This modular lesson material is intended for HPC software developers, computational scientists, research software engineers, and technical decision makers who want to understand the GPU programming ecosystem and identify the most suitable technologies for their applications and long-term software development goals.

By the end of this module, learners should be able to:

Understand when and why GPUs and other accelerators are beneficial for HPC applications (link to leaves in skill tree);
Grasp the core concepts of GPU programming, including parallel execution models, GPU architectures, memory management, and the challenges of developing efficient GPU-accelerated applications (link to leaves in skill tree);
Navigate the GPU software ecosystem by understanding the roles of different programming models, frameworks, libraries, and development tools (link to leaves in skill tree);
Evaluate and select appropriate GPU programming tools and technologies based on application requirements, performance objectives, portability needs, and existing software constraints (link to leaves in skill tree).

Credits and Licenses¶

Credit

Don’t forget to check out additional modular lesson materials from the EVITA platform. Please contact us if you want to reuse these course materials in your teaching. You can also join the XXX channel to share your experience and get more help from the community.

License

CC BY-SA for media and pedagogical material

Copyright © 2025, EVITA project, Cristian-Vasile Achim, Hicham Agueny, Andrey Alekseenko, Richard Darst, Karim Elgammal, Francesco Fiusco, Juan de Gracia, Johan Hellsvik, Jaro Hokkanen, Erik Holmström, Qiang Li, Wei Li, Daniel Medeiros, Ashwin Vishnu Mohanan, Tapish Narwal, Pedro Ojeda May, Yann Pfau-Kempf, Stephan Smuts, Stepas Toliautas, Apostolos Vasileiadis, Yonglei Wang, and Kjartan Thor Wikfeldt.

This material is released by the EVITA project, Cristian-Vasile Achim, Hicham Agueny, Andrey Alekseenko, Richard Darst, Karim Elgammal, Francesco Fiusco, Juan de Gracia, Johan Hellsvik, Jaro Hokkanen, Erik Holmström, Qiang Li, Wei Li, Daniel Medeiros, Ashwin Vishnu Mohanan, Tapish Narwal, Pedro Ojeda May, Yann Pfau-Kempf, Stephan Smuts, Stepas Toliautas, Apostolos Vasileiadis, Yonglei Wang, and Kjartan Thor Wikfeldt, under the Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).

Canonical URL: https://creativecommons.org/licenses/by-sa/4.0/

See the legal code

You are free to

Share — copy and redistribute the material in any medium or format for any purpose, even commercially.
Adapt — remix, transform, and build upon the material for any purpose, even commercially.
The licensor cannot revoke these freedoms as long as you follow the license terms.

Under the following terms

Attribution — You must give appropriate credit , provide a link to the license, and indicate if changes were made . You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.
ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.
No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.

Notices

You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation .

No warranties are given. The license may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.

This deed highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. You should carefully review all of the terms and conditions of the actual license before using the licensed material.

MIT for source code and code snippets

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.