Synopsys Executive: Nvidia Deal Will Help Partners ‘Solve Larger Problems Faster’

Synopsys executive Steve Pytel explains to CRN how channel partners will benefit from its new blockbuster alliance with Nvidia, which is expected to result in an overhaul of its products, including the analysis and simulation applications gained through its Ansys acquisition.

A Synopsys executive said the engineering software giant’s new blockbuster alliance with Nvidia will allow its channel partners to “solve larger problems faster” for customers because of its plan to significantly boost application performance with GPUs.

“Our customers want to solve problems faster, and they want to solve larger problems, and any technology that we deliver adds a value to that,” said Steve Pytel, vice president of product management at Synopsys, in an interview with CRN last week.

Announced last week, the “expanded, strategic partnership” between the two companies includes a $2 billion investment by Nvidia in Synopsys common stock as well as work to improve the performance and capabilities of the latter company’s software products.

Among the products set to receive an Nvidia-powered overhaul are the analysis and simulation applications Synopsys gained through its $35 billion acquisition this year of Ansys, which has traditionally relied on the channel partners to sell products.

“[Partners are] an extremely important piece of our business. We have a very significant channel,” said Pytel. “I can’t emphasize that enough.”

The executive said that “many if not almost all” of Synopsys applications take advantage of GPUs today to improve performance but added that use may be limited to certain aspects of each application, such as visualization.

One channel partner that has already seen the fruits of Ansys’ move to GPU acceleration is Burnsville, Minn.-based systems integrator Nor-Tech. Dominic Daninger, vice president of engineering at Nor-Tech, said the growth of GPU-bound workloads and the new capabilities they enable have opened new customer opportunities for his company.

“We’ve seen a lot of opportunities where [organizations], particularly universities, are interested in experimenting with AI,” he said. These opportunities usually start out small, according to Daninger, “but some evolve into good-sized opportunities.”

“And usually if we’re the ones that supplied the initial [deployment], we get a very good shot at the larger opportunities that come down the road a year or two later,” he added.

With Synopsys’ new Nvidia partnership, the company is making bigger changes to its applications under the hood to increase performance by orders of magnitudes. In the early work done so far, the two vendors claimed to speed up a computational fluid dynamics workload by 198 times and an AI-based physics simulation by more than 1,000 times.

“What we’re really looking to do is to take these software applications and make a massive improvement in performance,” Pytel said.

Synopsys plans to achieve this by using Nvidia’s CUDA-X software libraries, which are designed to accelerate specific application domains with its GPUs, along with its AI physics technologies, which includes its new Apollo family of open models.

The companies said they also plan to integrate Synopsys AgentEngineer technology within Nvidia’s agentic AI technology stack to “enable autonomous design capabilities for electronic design automation and simulation and analysis workflows.”

But Pytel said these changes, particularly the ones involving deeper utilization of the CUDA-X libraries, won’t happen overnight across Synopsys’ portfolio because of the work it will take to accelerate applications that were originally designed to run on CPUs.

“You don’t write code that was essentially an x86 instruction set and immediately realize the benefit of a GPU technology,” he said. “Some people will say you can port it, but sometimes the numerical algorithms themselves have to be rewritten in a way to support that change in silicon hardware architecture.”

Echoing the joint announcement by Synopsys and Nvidia last week, the executive noted that this partnership is not exclusive, which means the company will continue working with the likes of Intel and AMD to improve the way its software performs on their CPUs.

“[We’re] going to continue to make improvements on x86, but we’re going to accelerate what it means to use GPUs to get that next level of performance,” he said.

This kind of work is expected to result in closer collaboration between teams at Synopsys and Nvidia, according to Pytel. He added that the partnership could also push both companies to dedicate more employees over time to the integration work.

“Could an outcome be that there’s more people working on it? Absolutely that could be an outcome,” he said. “But the first step is getting alignment with the teams.”

While there is much work ahead for Synopsys and Nvidia, Pytel said channel partners have already reaped from the benefits of recent collaboration efforts between the two vendors.

As an example, he pointed to CADFEM, a German solution provider in the Ansys Apex partner program. The partner contributed to the design of a simulation of an assembly line to serve as a digital twin for packaging and bottling company Krones. The goal of the simulation-driven framework is to optimize manufacturing processes in real time.

Presented alongside Synopsys at the Microsoft Ignite event last month, CADFEM customized the solver settings within the Ansys Fluent computational fluid dynamics application to help Krones take full advantage of the baked-in GPU acceleration.

This contributed to significant speed-up for the application, from three to four hours to less than five minutes to run different scenarios in the simulation, according to Synopsys.

“This is just more opportunity for our resellers to essentially go and be innovative with our customers,” Pytel said.