Cradle computational flow dynamics (CFD) software was first released in 1984. The company services over 7,000 licenses and runs more than 500 analyses every year for its customers. Electronics companies use Cradle to design cooling systems, and automotive companies use it to analyze aerodynamics and powertrain performance.
Ship builders can test body shape and performance while optimizing vessel drag and propeller efficiency. With an eye on saving energy and lowering CO 2 emissions, Cradle pushes the idea that its thermal and fluid analysis tools can save companies engineering time and money by building virtual prototypes and running concurrent simulations that won’t require physical tests.
scSTREAM helps architectural designers model heat flow. (Image courtesy of Cradle CFD.)
Cradle’s parent companies Hexagon and MSC software recently announced that the CFD tool would partner with Fujitsu, creators of the the supercomputer Fugaku, to add additional computing power. Fugaku, developed jointly by RIKEN, is one of the world’s most powerful computing systems, but its aim is to work toward making the world a better place. Fujitsu’s purpose is “to make the world more sustainable by building trust in society through innovation.”
Fujitsu started development of the supercomputer way back in 2014, but the pandemic shifted its focus. The large numbers of engineers and programmers working from home showed that the world is more reliant than ever on computing power. Moving massive chunks of the workforce out of the office seemed like an almost impossible task 10 years ago, but when stay-at-home orders forced people to stay inside, a lot of work continued. Fujitsu sees this style of work as a permanent change internally, and now seeks to enable the rest of the world to navigate this work life shift.
Beyond Fujitsu’s commitment to getting more power to people working from home, the company has “ Priority Issues” built around the goal of making the world a better place. The first priority is a commitment to promoting health and longevity with a focus on a drug discovery infrastructure and supporting both personalized and preventative medicine. Other priorities share the same ambitious goals, from mitigating disasters to developing renewable energy sources.
The Supercomputer Meets the Super Software
By leaning into the supercomputer, Hexagon hopes to “analyse all the complexities of reality with less than half the energy use and at a fraction of the cost of traditional simulation methods.” The company’s Manufacturing Intelligence division has found success in developing vehicles and aircraft when simulating high levels of iterations to optimize designs. Hexagon estimates that in past projects up to 90 percent of an engineer’s job might have been dedicated to readying the model for analysis.
CFD simulations require large amounts of computer power for even one iteration, and one of the ways to work within the constraints is model simplification. Geometry sometimes needs to be simplified, with features removed or complex interfaces between parts turned into a single subassembly. Engineers need to decide which simplifications will give the processor more time to work the simulation and which are too critical for the simulation and can’t be removed without changing the results of the study.
More sophisticated designs are requiring more iterations and more realistic modeling, but working to meet customer needs while performing a true and meaningful test can be a challenge. Hexagon hopes that giving Cradle engineers access to Fugaku’s PRIMEHPC series supercomputer will open a new field of results.
What Can CFD Specialists Do with Fugaku?
Taking away the need to simplify designs before analysis occurs will free up engineering time and provide access to more iterations and more detailed results. The Fugaku computing platforms require about one-third of the energy that typical Cradle CFD computers are currently using. Hexagon sees immediate possible uses in all its customer industries but large opportunities in vehicle development and infrastructure. The electrification of vehicles is full of unknowns; the ability to create more iterations for optimization might produce more innovation. Charging infrastructure and the changes required for roads and cities can also benefit from a wider range of possibilities and combinations.
The rigid and elastic bodies of two-way fluid structure interaction studies. (Image courtesy of Cradle CFD.)
Roger Assaker, president of Design & Engineering in Hexagon’s Manufacturing Intelligence division, said “Simulation holds the key to innovations in aerospace and eMobility. Advances such as the low-power Fugaku supercomputing architecture are one of the ways we can tap into these insights without costing the Earth, and I am delighted by what our Cradle CFD team and our partners have achieved.”
Bringing engineering teams together from the two companies has already yielded promising results. Hexagon engineers were able to produce an automotive simulation that they claim could only be possible when using the Fugaku computing power. The model used 960 cores to perform Reynolds-averaged Navier–Stokes (RANS) equations until steady state, over more than 70 million elements.
Another test revolved around an aircraft study performed with the RIKEN Center for Computational Science (CCS) and the High Performance Computing Infrastructure (HPCI) groups. The study involved turbulence and its effects on an aircraft’s structural integrity. Because of current model simplification, the aircraft designers were concerned that they might not understand the full effects of turbulence on an airplane’s wings.
The study took an aircraft and simulated a transonic compressible fluid boundary. The full model had 230 million elements and used 192,000 computing cores to test 4,000 nodes. Message passing interface (MPI) standards and four threads in the OpenMP API helped the study to run efficiently. The user report from the study is available for review in Japanese. Next generation aircraft will need more complex models and more iterations to complete these types of studies, especially as supersonic and hypersonic shockwaves are brought into the mix.
What Happens Next?
Using a supercomputer to solve any problem will work, but part of sustainable management of resources means that the projects Hexagon chooses for this partnership will need to be vetted. The process of using Fujitsu’s resources and Cradle’s software has been proven and now the task of identifying the best projects will begin. Fugaku’s development was sponsored in part by the Japanese government, so several layers of regulations exist. Fujitsu has a document stating the Regulations for Use, guidelines for how users can store their data on the supercomputer, as well as a separate document discussing how users’ personal data can be stored. HPCI resources are available to public users defined as “Individuals who belong to universities, research institutes, or companies are eligible to use the HPCI system after the selection process,” but a lengthy process is in place to gain approval, and the organization only opens its Call for Proposals once a year in the fall.
Tomohiro Irie, director of R&D for Cradle CFD, said, “By using the efficient computing power of Fugaku with our simulation tools, we will encourage users to simulate phenomena that simply weren’t feasible before due to the computation time and cost. Today we have simulated with 192,000 cores, but this is only the beginning—because Cradle CFD is used in diverse applications, I expect that these technical developments will contribute to making the power of Fugaku more accessible for general use, bringing huge freedom and improved insights to engineering teams solving tomorrow’s problems today.”