The last time we were in a setting like this, Brad Rothenberg found himself talking about the future.
He’d been buoyed by the immediate aptitude of a teenage intern, whose successful tinkering with the company’s computational design tools served as strong validation for the nTop CEO’s professional focus. And he was allowing himself to wonder.
We’re at Formnext as we pick back up on two themes from our previous conversation at RAPID + TCT: the era of computational design and the importance of collaboration. The two serve each other. The former won’t be enabled without the latter, and there is little point in the latter if you’re not out to achieve something as grand as the former.
What nTop sees in front of it is opportunity. And as it chips away at perception and understanding, it is beginning to usher in what was first uttered in an nTop blog post published in June. A new era in design.
“Computational design is primetime now,” he told TCT at Formnext. “It’s ready for mainstream and it’s becoming mainstream.”
nTop has had a decent amount of pick-up since it came to market six years ago, but sees the hundreds of companies using its technology today as only the start. What lies ahead of the company is the potential to make computational the standard way of designing products. And nTop, Rothenberg believes, can serve as the foundational technology to speed up engineering. But it won’t be if it doesn’t take steps to reduce bottlenecks and harness the capabilities of others.
“In the new world of computational design, compute is the biggest bottleneck to speed, whereas in the old world, the bottleneck was how fast your best CAD person could model something,” Rothenberg said. “If you make a CAD system 100 times faster to solve the geometry problems, it’s still fragile, and you still need to manually draw up the changes. The limiting factor in computational design is compute. And so, we want to ensure that compute is never a bottleneck to computational design.”
If you wanted to know how nTop came to partner with NVIDIA earlier this year, that’s it. The partnership sees nTop integrate its computational design capabilities with NVIDIA’s OptiX rendering framework and Omniverse technologies, with the software giant also investing into nTop.
Together, they are working to train up AI surrogate models on top of nTop to achieve inverse design. Typically, an nTop user would set up parameters like thickness, height, print process and overhang, change the parameters, and then explore the outcomes. But the partners are attempting to generate AI surrogate models so users can start with the desire outcome and then receive the required inputs.
This is one facet of the burgeoning collaboration between the two companies. And it is just one collaboration in a growing number of cooperative projects that nTop is undertaking. Rothenberg points to the work being carried out with cloudfluid as another example.
Get your FREE print subscription to TCT Magazine.
Exhibit at the UK's definitive and most influential 3D printing and additive manufacturing event, TCT 3Sixty.
cloudfluid is a computational fluid dynamics solver that which is said to work well with nTop’s native implicit format. Their partnership means users no longer have to export a mesh to simulate their model, with cloudfluid able to package a ‘super lightweight’ nTop *.implicit model and use a cloud-based GPU solver to compute the results quicker. Users only need to set up the boundary conditions they want and click a button to kickstart the process.
“As a customer, you want to use best-in-class tools and the best-in-class tools all have to speak the same language,” Rothenberg said. “If you have to translate from one language to another, and certain words in one language are not translatable, that becomes really difficult.”
It is partnerships like this that will pave the way for the era of computational design. Rothenberg notes that what users need to fully take advantage of the tools nTop has brought to market are tight integrations between this ‘new process’ and the rest of the workflow.’
To that end, the partnership with Materialise Magics announced in the summer hasn’t been a one-and-done alliance, the pair have continued to build on the relationship. In June, an early access program with leaders from the aerospace, automotive, Formula 1 and medtech industries was launched, and last month it was announced that Additive Industries, Stratasys and Renishaw are now involved in the development of Materialise’s Next-Gen Build Processors.
nTop’s implicit modelling kernel has been an optional add-on to all new Next-Gen Build Processors for six months, with Rothenberg suggesting it will do what nTop is striving to do: reducing the biggest bottleneck to faster design.
That bottleneck, as Rothenberg describes it, is how users get from a digital model to a physical part. The company’s computational design software is built on a new data model – a signed distance field-based implicit model – which runs natively in GPUs and allows users to make changes to a model ‘an order of magnitude faster’ than is possible with CPU-based single core models. But even with these quick physics-driven alterations that allow users to hone in on the right design available, they would still have to go back to a legacy format when meshing and exporting.
“There was a gigantic lag of having to clean up a mesh, export a mesh, and some parts couldn’t even mesh at all,” Rothenberg explained. “Imagine a heat exchanger the size of a table with hundreds of thousands of pores and holes going through it, all interlocking. It's not meshable. You make a mesh and the mesh would be 60 gigabytes. You can't open that in Magics, you can't manufacture it.”
nTop, as Rothenberg puts it, has released a ‘massive power’ to the world. But to make the application of that power practical, it has recognised the challenge would be insurmountable if it was to go it alone.
“What's so significant today is with Materialise, with the initial announcement of Materialise incorporating implicit interop, it unlocks that power for a subset of their users. They're saying all of their tool chain is going to support implicit and that's massive, because it means you could use nTop and explore, ‘what if I print on this machine?’ ‘What if I print on this machine?’ ‘What if I print on this machine and instantly get the feedback from that?’”
These partnerships then are about eliminating the what ifs, or at least providing answers to them. Since launching its computational design tools, the customer feedback has been positive, but often caveated. Manufacturers have been able to save time and generate greater design complexity in product development, only to be slowed down when trying to make the leap from digital to physical, from idea to reality.
That is relatable for nTop because it too is trying to make such a stride. Currently, the era of computational design is just words on a webpage. It’s still a concept rather than an actuality.
But for nTop, it’s a north star. It is what motivates the company day-to-day and what is driving it to align with other industry-leading software tools. Is the idea of a computational design era a sure thing? For nTop it is.
“We’re far from there, but absolutely,” finished Rothenberg. “That’s why I do what I do.”
