About ARC4CFD
Computational Fluid Dynamics (CFD) is a field of computational physics that has a very high utilization of modern advanced research computing (ARC) resources. CFD codes solve the discretized form of the governing equations for fluid motion, which results in a high-degree-of-freedom mathematical problem. The spatial and temporal resolution required to solve modern problems makes CFD particularly well suited to take advantage of the full benefits of large-scale, parallelized high-performance computing on ARC infrastructure. This course is designed to help students with a cursory knowledge of fluid dynamics and CFD to bridge the gap to effectively leverage modern ARC resources.
Teaching philosophy
ARC4CFD is a web-based, fully open-source, asynchronous, online course. The multimodal course content includes written text, videos, interactive quizzes, and hands-on examples covering a wide range of topics relevant to CFD on ARC. This course uses an entirely open-source CFD toolchain from meshing all the way to solving, and visualization, in order to facilitate sharing. The course is developed in order to: 1) develop of a systematic approach for the effective utilization of CFD usage on ARC systems, 2) provide a high-level, theoretical understanding of the main concepts in using CFD on HPC, and 3) provide hands-on examples for the learners to put into practice these concepts. The course is divided into three sections, which are made up of a number of classes each with clearly defined learning outcomes. These sections are structured in a way for the reader to first develop a foundational understanding of high-performance computing (section 1), translate those concepts to specific challenges in CFD and develop a systematic approach towards a CFD workflow (section 2), and then to effectively manage the research data throughout the entire workflow (section 3). The course is built around the end-user of the CFD tools, not experts, nor code developers. In other words, the focus lies in providing the end-user with the knowledge necessary to setup a workflow and strategies to effectively run CFD simulations on HPC systems.
Learning outcomes
At the end of this course, the learner should be able to:
- Define the main concepts in parallel and high performance computing
- Conduct an a priori estimate of the computational cost of a CFD simulation
- Explain the impact of modelling assumptions on HPC cost
- Optimize simulation parameters of a CFD problem for HPC
- Develop a research data management strategy for a CFD workflow
It’s expected that the learners have a basic understanding of fluid dynamics and computational fluid dynamics as this course is designed to bridge the knowledge gap between basic CFD knowledge and the utilization of these tools on modern HPC systems.
Target audience
The target audience for this course is:
- new graduate students in computational physics or engineering,
- experimentalists and theoreticians complementing their work with numerical simulations,
- undergraduate students on student design teams interested in leveraging CFD with HPC.
This course was developed in the Multi-Physics Interaction Lab at the University of Waterloo, Canada with the financial support of Compute Ontario.