FIFTY2

Innovation corner

There is always one more shot to solve the unsolved problem. Tinkering around, entering the unknown and starting over again is our approach to push the boundaries and create next level innovations. Stay tuned for PreonLab updates, new researches, groundbreaking innovation and upcoming events.

March 15, 2024
Siddharth Marathe
With every PreonLab update, we aim to continuously enhance your simulation experience by increasing efficiency and reducing the memory footprint. While simulation on CPU is still the bread and butter for CFD, it is undeniable that GPUs can provide a significant performance boost towards reducing computation time. Nevertheless, one advantage CPUs generally have over GPUs is a larger memory space. Due to the limited memory of single GPU cards, simulating very large scenes with a lot of particles can be quite challenging. In addition, importing large tensor fields like airflows can also occupy a lot of precious memory space on the graphic card. While PreonLab can cleverly resample such airflows to fit on single GPU hardware, there will always come a point when sacrificing accuracy will be inevitable. One logical solution is making use of state-of-the-art GPU hardware that can accommodate large simulation scenes. Professional GPU cards like Nvidia’s H100 GPUs can already offer memory space up to 80 GB. However, does this mean that GPU simulations are only possible through the acquisition of larger and larger professional GPU cards? And what about scenes which might require even more memory space than the latest hardware available?
February 29, 2024
Jan Viher, Siddharth Marathe and Markus Ihmsen
In the field of Computational Fluid Dynamics, the pursuit of simulation efficiency and accuracy is the name of the game. These two seemingly contradictory goals often appear at opposite ends of the spectrum. But is this always the case? Is it possible to have both efficiency and accuracy without making sacrifices? We believe it is. And we’re not referring to hardware efficiency, a topic extensively covered in our previous articles (see here and here). What we mean is state-of-the-art software development which is in the DNA of the FIFTY2 team. What we want to talk about in detail is one of our most prominent features available only in PreonLab. We named it Continuous Particle Size, or in short, CPS.
We are delighted to announce our new release – PreonLab 6.1. It is loaded with new features and improvements, enabling new engineering possibilities and applications. Empowered by the new GPU capabilities, PreonLab 6.1 takes another leap forward in efficient computing, delivering results even quicker. Read on to find out more about: Enhanced GPU performance: PreonLab’s GPU implementation just got a significant boost. Not only does PreonLab 6.1 support multi-GPU computing now, but it also enables continuous particle size (CPS), dynamic sampling, and adaptive sampling on GPU, making your simulations even more efficient. Airflow and Car Suspension Model (CSM): PreonLab 6.1 also enables airflow import and CSM support for simulations on GPU. These are some of the key application enablers making it possible to simulate various use cases also on GPU. Snow Model on GPU: Snow modeling has been an important part of existing PreonLab capabilities. We are thrilled to announce that our snow model is now also available on the GPU platform. Enhanced Thermodynamics: Convective boundary condition has been added to the list of available boundary conditions.  Convective Boundary Conditions can represent a more physical heat transfer, that predominantly occurs due to convective heat transport at fluid-fluid or fluid-solid interfaces. It is designed to conveniently represent natural or forced convection of heated solid bodies or fluids, in applications like heat exchangers, heat sinks, and even e-motors. This is just a selection of new features and improvements. Check out the changelog to learn about all the changes.  Make sure to follow us on LinkedIn so that you don’t miss new videos, case studies and updates!
December 21, 2023
Jens Cornelis, Andreas Henne and Loïc Wendling
Preoneers rely on the fact that with every new release, PreonLab aims to enhance your simulation experience by increasing efficiency and reducing the memory footprint for simulations. We firmly hold this at the core of our vision towards developing the ultimate simulation tool. We are glad to announce that version 6.1 takes a great leap towards realizing this goal as it extends GPU support for many PreonLab features and introduces multi-GPU capabilities.
September 05, 2023
Siddharth Marathe
As simulation engineers, we want to get our ideas and points across to stakeholders, managers, co-workers from other departments, and customers as efficiently as possible. Typically, we back up our arguments by generating plots for statistics, which we put together to analyze our simulation results. We know exactly what all those numbers and multiple graphs and diagrams on every PowerPoint slide mean, and what conclusions need to be drawn from them to improve the product’s design. However, this is not always the case for everyone else, who is looking at the presentation. It is possible that some viewers might even be viewing the results and concepts for the first time, without much in-depth technical knowledge. Furthermore, even the most experienced engineer or team member might interpret a simulation result inaccurately, if all they have time for is a quick glance at the data. So, when it comes to communicating these results effectively, the message needs to be conveyed clearly – ideally in an appealing and easy-to-understand manner. This can be quite challenging since there are many small details that matter, but we also do not want the viewer to lose focus from the big-picture point of the discussion. Showing graphs and numbers is important, but on its own this can be rather dry.
August 23, 2023
Jan Viher
“No wind blows in favour of a ship without direction.” This wisdom resonates deeply with our journey at FIFTY2, as it emphasizes the importance of having a clear vision. From day one, our direction at FIFTY2 has been unmistakable – to develop the ULTIMATE simulation tool. This is the story of how it is done.
July 28, 2023
Andreas Henne
When we started looking into CUDA support for PreonLab I didn’t really have any experience programming for GPUs. Getting from the first experiments in 2021 to production code has been a long learning process. For CUDA, there is no shortage of useful articles, documentation and examples (in fact this was one of the reasons to pick CUDA as a platform). Nevertheless, I thought it might be interesting to go over the most important lessons from my perspective as a developer used to x86-based CPUs, so here we go!
“Why can’t I run PreonLab on my GPU?” We often heard this question and didn’t have a good answer. But behind the scenes, we have been experimenting with GPUs for quite some time. We are excited to announce that we are finally ready. PreonLab 6.0 is here, and it changes everything. This is how: GPU Support using Nvidia CUDA: PreonLab 6.0 delivers the greatest leap in efficiency of any update so far. For a wide range of applications, a single GPU can deliver up to 6x faster performance compared to a state-of-the-art multi-socket CPU system. If you have the chance, please try PreonLab on GPU – you will not want to go back. A New Engine for PreonLab: We believe that the future of PreonLab is multi-platform. CUDA is an amazing platform, but it is not the only one we wish to explore in the future. To this end, we have rewritten our simulation core with platform independence in mind. Expect more on this in upcoming releases. Lateral Adhesion: The new experimental “Lateral Adhesion” option models droplet sticking and runoff behavior more accurately. This is important for applications such as tailgate runoff simulation. Improved User Interface: Various changes to the graphical user interface such as the new welcome screen make using PreonLab more enjoyable than ever. This is just a selection of new features and improvements. Check out the changelog to learn about all the changes.
June 29, 2023
Siddharth Marathe and Andreas Henne
“Time is money”. In the world of simulation this means being able to perform reliable simulations, faster. Throughout its development, PreonLab has already introduced game-changing approaches, like its unique implicit IISPH formulation (learn more about it here) and advanced adaptive refinement and coarsening with Continuous Particle Size (read about it here) for efficient simulation on CPU hardware. Over the past decade, the development of more advanced GPUs has gathered up great speed and it is likely that this trend will continue in the future. Behind the scenes, we too have been experimenting with GPUs for some time now, to fully leverage their potential for PreonLab. Now, PreonLab’s code has been ported to run on GPUs making use of Nvidia’s proprietary CUDA API to push the speed-barrier for particle-based simulation even further.
June 22, 2023
Alexander Mayer
Slamming is a term that is often used in a maritime context to describe the sudden impact of the ship hull on a water surface. It leads to pressure spikes on the hull along with rapid and intricate deformations of the fluid surface. Physical testing, especially for large-scale applications in the maritime world, is not only time-consuming and expensive but often even unfeasible. Computational fluid dynamics can complement real-world experiments and hence reduce costs as well as accelerate development. However, simulating the water entry of solid bodies is no easy task. Using traditional grid-based methods requires periodical remeshing due to the moving geometry and discretizing the entire simulation domain. Particle-based simulation methods on the other hand can generate insights without these inconveniences, saving valuable computation time. This article aims to show how PreonLab can be used to simulate the water entry of free-falling rigid bodies.