How AI Will Change Chip Design

end Moore’s law Looming on the horizon. Engineers and designers can do a lot about this Miniaturization of transistors and Pack as many of them as possible into chips. So they turn to other methods Chip designand integrating technologies such as artificial intelligence into the process.

Samsung, for example, is Adding artificial intelligence to its memory chips To enable processing in memory, thus saving and accelerating energy Machine learning. Speaking of speed, Google’s TPU V4 AI chip has it Double the processing power Compared with its previous version.

But artificial intelligence holds more promise and potential for the semiconductor industry. To better understand how AI is poised to revolutionize chip design, we spoke with Heather goreSenior Product Manager for mathworksMatlab platform.

How is AI currently being used to design the next generation of chips?

Heather Gore: AI is an important technology because it is involved in most parts of the cycle, including the design and manufacturing process. There are a lot of interesting applications here, even in general process engineering where we want to improve things. I think defect detection is a big deal at all stages of the process, especially in manufacturing. But even thinking ahead in the design process, [AI now plays a significant role] When you’re designing the light and the sensors and all the different components. There’s a lot of anomaly detection and mitigation that you really want to take into account.

Heather goremathworks

Then, when thinking about the logistics models you see in any industry, there is always planned downtime that you want to mitigate; But you also end up taking unplanned downtime. So, if we look back at that historical data of moments where maybe something took a little bit longer to manufacture than expected, you can look at all that data and use AI to try to pinpoint the immediate cause or to see something that might be showing up even in the processing and design stages. We often think of AI as a predictive tool, or as a robot doing something, but often times you get a lot of knowledge from data through AI.

What are the benefits of using AI to design slides?

Gore: Historically, we’ve seen a lot of physics-based modeling, which is a very intensive process. We want to do a Discounted order formwhere instead of solving such a computationally expensive and exhaustive model, we can do something a little cheaper. You can create an alternative model, so to speak, to this physics-based model, use the data, and then do whatever you want Parameter scansyour improvements, your Monte Carlo simulation Using the alternative model. This takes much less computational time than solving physics-based equations directly. So, we see this benefit in many ways, including the efficiency and economy that comes from the rapid iteration of experiments and simulations that will really help with the design.

So it’s kind of like you have a digital twin?

Gore: exactly. That’s pretty much what people do, where you have a physical system model and experimental data. Then, in conjunction, you have this other model that you can tweak and fine-tune and try different parameters and experiments that allow you to go through all those different situations and come up with a better design in the end.

So, it would be more efficient and, as you said, cheaper?

Gore: Yes, of course. Especially in the experimentation and design phases, where you’re trying different things. Obviously this will result in significant cost savings if you are already manufacturing and producing [the chips]. You want to simulate, test and experiment as much as possible without making something using actual process engineering.

We’ve talked about the benefits. What about defects?

Gore: the [AI-based experimental models] They tend not to be as accurate as physics-based models. Of course, that’s why you run so many simulations and parameter scans. But that’s also the benefit of having this digital twin, because you can keep that in mind – it’s not going to be as accurate as this exact model that we’ve developed over the years.

Both chip design and manufacturing are system-intensive; You have to think about every little part. This can be a real challenge. It’s a case where you may have models to predict something and different parts of it, but you still need to bring it all together.

One of the other things to think about as well is that you need the data to build the models. You have to integrate data from all different types of sensors and different types of teams, and that adds to the challenge.

How can engineers use AI to better prepare and extract insights from device or sensor data?

Gore: We always think about using AI to predict something or do some robot task, but you can use AI to come up with patterns and pick out things that you might not have noticed before yourself. People will use AI when they have high-frequency data coming from many different sensors, and oftentimes it’s useful to explore the frequency domain and things like data synchronization or resampling. This can be a real challenge if you’re not sure where to start.

One of the things I would say is to use the tools available. There is a large community of people working on this stuff, and you can find a lot of examples [of applications and techniques] on github or Matlab Centralwhere people shared great examples, even small apps they’ve created. I think a lot of us are buried in data and aren’t sure what to do with it, so definitely take advantage of what’s already out there in the community. You can explore and see what makes sense for you, balancing domain knowledge with the insight you get from tools and AI.

What should engineers and designers consider?en Using artificial intelligence to design chips?

Gore: Think about the problems you are trying to solve or the ideas you might hope to find, and try to be clear about that. Think about all the different components, and document and test each of those different parts. Consider all the people involved, explain and deliver in a way that makes sense to the whole team.

How do you think AI will impact the jobs of chip designers?

Gore: This will free up a lot of human capital for more advanced tasks. We can use AI to reduce waste, improve materials, improve design, but then you still have that human involved when it comes to decision making. I think it’s a great example of people and technology working hand in hand. It’s also an industry where everyone involved – even in manufacturing – needs to have a certain level of understanding of what’s going on, so it’s a great industry for AI development because of the way we test things and how we think about them before we put them in place. them on the slide.

How do you envision the future of artificial intelligence and chip design?

Gore: It’s very much based on this human element, which is getting people into the process and having this interpretable model. We can do many things using the mathematical details of modeling, but it’s about how people use it, and how everyone in the process understands and applies it. Communicating and involving people of all skill levels in the process will be really important. We’re going to see fewer of those hyper-accurate predictions and more transparency in information and sharing and that digital twin – not just using AI but also using our human knowledge and all the work that many people have done over the years.