AI PCB Design: How Generative AI Takes Us From Constraints To Possibilities

Generative artificial intelligence (AI) represents the next great step forward in PCB design.

This is, of course, what you might expect me to say. 2023 has been a year dominated by the rise of generative AI, with large language models (LLMs) as the unquestionable poster child. It’s easy to see why; LLMs seemingly fulfill the promise of sentience, while in reality, they’re simply very good at inferring context from both historic and real-time data.

But while the term “generative AI” may only recently have entered the global lexicon, Cadence has been applying generative AI to design for a fair while longer. We’ve called it generative design, but it’s essentially the same thing—an AI capable of taking context and a set of constraints and applying human-like reasoning in its response.

Cadence customers are already reaping the benefits of generative AI within our Joint Enterprise Data and AI (JedAI) Platform. Chip designers are realizing Cadence Cerebrus AI to design chips that are faster, cheaper, and more energy efficient.

Now, we’re bringing this generative AI approach to an area of EDA that has traditionally been highly manual—PCB placement and routing.

AI PCB design makes sense

When PCB layout tools were first introduced, they weren’t much better for routing traces than using an etch-a-sketch. Today, constraint-driven automation approaches have reduced the time required to perform certain design tasks—yet PCB placement and routing (P&R) remains a relatively manual balancing act, requiring the human operator to consider signal integrity, power plane delivery, thermal management, and more.

This process has only become more complex as the industry has embraced multi-layer, 3D stacking, and heterogeneous integration. Inevitably, this has led to designers sacrificing certain design aspects for the sake of others, adopting more PCB layers than needed or taking a “good enough” approach to avoid extended time to market.

We know that AI outperforms humans at intensely computational design tasks. Generative AI presents a compelling solution to the growing complexity of today’s and tomorrow’s PCBs—which is why we put generative AI front and center in our Allegro X AI product, announced earlier this year and soon to be available to Cadence customers.

Allegro X AI

The foremost advantage of employing generative AI in PCB design lies in its unmatched capacity to explore a vast design space and identify alternative solutions with speed and efficiency.

That’s not to say that given infinite time and resources, a human might not reach the same conclusions. Generative AI isn’t necessarily smarter than us—but it is a whole lot faster at something upon which competition and time-to-market impose ever greater limits.

Allegro X AI flips the PCB design process on its head. Rather than present the operator with a blank canvas, it will take a list of components and constraints that need to be satisfied in the end result and sift through a plethora of design possibilities, encompassing varied placement and routing options. This is hugely powerful for hardware engineers focused on design space exploration (DSE). This has long been par for the course in IC design yet it has more recently become critical to PCB due to the fact that today’s IC complexity doesn’t reduce when it gets onto the PCB—it increases.

If you’ve engaged in anything like an extensive conversation with an LLM, you’ll know that you can challenge it to regenerate its output based on additional factors. That works here, too; a designer could adjust a constraint to see whether Allegro X AI can generate a viable output in 8 layers instead of 10, for example. When every layer adds 10% to the cost of the overall board, that’s a significant cost saving.

This exhaustive exploration of large-scale designs enables designers to optimize routing for signal integrity, power distribution, and thermal management. All while reducing design turnaround from days to hours if not minutes. Faster design iterations mean quicker prototyping and product development, ultimately leading to more boards produced per year and reduced time-to-market for end products.

Rules are rules

Design creativity is another of the foundational differences between PCB and IC design. In IC, we are still largely limited to Manhattan routing. But with PCB, we have so many more degrees of design freedom that classical algorithms can’t solve without massively compromising the design and its wire length. That’s where human creativity has traditionally had to step in—at the expense of days of trial and error in placement.

However, it’s important to understand that this isn’t Cadence replacing traditional compute algorithms and automation approaches with AI. We remain as committed to accuracy and “correct by construction” as we’ve ever been, and while Allegro X AI is trained on extensive real-world datasets of successful and failed designs, we don’t use that data to determine correctness.

The result is a combination of physics-based design algorithms with generative AI, capable of using those algorithms as heuristics to ensure that its output is correct while drawing on the collective wisdom of the past to leverage layout best practices that may not be explicitly present in traditional algorithms.

It’s this human creativity that generative AI emulates so well. The tangible result of adding one to the other? Design excellence, greater competitiveness, and significantly reduced time to market. Everything Cadence has ever stood for—now realized more than ever thanks to generative AI.