If only the harried engineer were simply a meme. But the fact is, engineers consistently report being under pressure to do more with fewer resources. Engineering talent shortages continue making headlines, and the job is getting increasingly complex and bigger in scope. While the slowing of Moore’s law is creating limitations on some semiconductor advancements, our digital world continues demanding much more from our electronics.
How can engineers deliver the complex silicon chips today required to tackle the challenges of tomorrow?
Now, there’s a proven solution that enhances engineering productivity and silicon quality of results—even making possible what has previously been impossible to accomplish. All of this comes, incidentally, from one of the very technologies that is driving greater chip complexity: artificial intelligence (AI).
Engineering ingenuity has led to advancements like AI-powered chatbots, surgery-performing robotics, and self-driving cars. It has also produced solutions that offload repetitive chip design, verification, and testing tasks, allowing engineers to focus on what they do best: innovate.
With the award-winning Synopsys DSO.ai™ AI application for chip design leading the way, Synopsys has unveiled the industry’s first full-stack, AI-driven electronic design automation (EDA) design suite, with solutions for functional verification (Synopsys VSO.ai) and silicon test (Synopsys TSO.ai) now available and more capabilities coming down the road. Customers with early access to the Synopsys.ai technology are reporting impressive results, from 10x improvement in reducing functional coverage holes to up to 30% increase in IP verification productivity. Meanwhile, DSO.ai has recently notched its first 100 production tape-outs, a significant milestone that marks AI’s ascent into the semiconductor mainstream.
In a global economy currently shaped by uncertainties, the semiconductor industry has grown even more competitive. These factors have increased the need for even better results for semiconductor power, performance, and area (PPA). Those who can optimize these metrics faster and at lower cost stand a better chance of rising to the top. As companies adjust to today’s economic realities, there are underlying staffing pressures, with projections of scarcities in semiconductor engineering talent that could hamper progression in the industry. According to the Semiconductor Industry Association, the U.S. semiconductor design industry could face a 23,000 shortfall of semiconductor engineers by 2030.
An AI-driven EDA design suite offers a way forward, complementing and enhancing the work of engineering teams. According to Deloitte Global, the world’s top semiconductor companies will spend US$300 million on internal and third-party AI tools to design chips this year, a spend that’s expected to grow 20% each year over the next four years. The analyst firm notes in an article, “AI design tools are enabling chipmakers to push the boundaries of Moore’s law, save time and money, alleviate the talent shortage, and even drag older chip designs into the modern era. At the same time, these tools can increase supply chain security and help mitigate the next chip shortage.”
In every design, there are time-consuming tasks that are iterative and somewhat redundant, yet necessary for the quality of the chip. AI technologies can take on (and enhance results of) repetitive tasks such as design space exploration, verification coverage, and test pattern generation. This frees up a substantial amount of time for engineering teams, allowing these experts to focus on value-added tasks such as differentiating their products and quickly creating new features or derivative designs. It also allows teams to meet aggressive time-to-market targets and to accommodate more projects with their existing resources. Engineers who are new and seasoned veterans alike can benefit from AI, as the technology provides faster ramp up time for the new engineer and enables the veteran to achieve better quality of results, faster. For example, on a given project, Synopsys.ai can be tasked with handling repetitive work and acting on iterative tool results while engineers oversee chip quality, together generating better, faster results than what would be possible by humans alone.
Designers are facing a multitude of challenges, from increasing design complexity to advanced process node requirements and shrinking power budgets across multiple verticals including CPUs, mobile, GPUs, automotive, and AI. AI chips, for example, need to be architected with many compute units to process sophisticated algorithms and huge amounts of data, whether in edge devices or in servers nestled inside data centers. Optimizing such complex designs for PPA and ensuring that the chip will perform as intended is a huge undertaking—one that is becoming increasingly challenging for designers.
This is where a solution like Synopsys.ai changes the landscape. Let’s take a closer look at each of the available components of Synopsys.ai.
DSO.ai is a disruptive application of AI in chip design for implementation. It can autonomously search the PPA design space for optimal solutions. By applying AI to chip design workflows, DSO.ai facilitates a massive scaling in the exploration of choices which was not feasible before.
The approach creates an opportunity for the technology to continuously build on the training data and apply what it has learned to, ultimately, accelerate tape-outs and achieve PPA targets. Another key advantage of AI is its support of reuse: the retained learnings gained for one project can be utilized for future projects, bringing greater efficiency into the design process. Since the solution is available in the cloud, customers can take advantage of greater flexibility, scalability, and elasticity for massive workloads. Users have reported productivity enhancements of more than 3x, power reductions of up to 15%, and substantial die size reductions. Figure 1 shares results from the application of DSO.ai in a high-performance data center CPU.
VSO.ai helps verification engineers reach coverage closure targets faster and find more bugs. The number of design state spaces in which a digital design can operate is nearly infinite, making it virtually impossible for humans to check each of these spaces to validate that the design will function as intended. The regression process could run for days, eating up compute resources through thousands of tests. Often, the last mile closure ends up being very labor intensive, with manual analysis on huge amounts of data limited in yielding actionable insights.
VSO.ai revitalizes this process, examining the RTL to infer coverage while also highlighting areas where coverage is needed, saving substantial time and ensuring a high ROI on the tests. Renesas reports achieving up to 10x improvement in reducing functional coverage holes and up to 30% increase in IP verification productivity using AI-driven verification with Synopsys VCS® functional verification solution, a part of Synopsys.ai.
Growing design complexity and size also weigh down the silicon test process. There are three key metrics to consider when evaluating the results from an automatic test pattern generation (ATPG) tool:
Traditionally, optimizing for one of these metrics (typically by hand) negatively impacts the others. Someone who is new to ATPG may not have a strong sense of how to tweak the tool to generate the desired program results.
Conversely, someone with a lot of experience may have biases that cause him/her to set up the tool to achieve a certain result, which may not prove optimal for a new design. TSO.ai eliminates the tradeoffs, running multiple copies of Synopsys TestMAX™ ATPG advanced test solution to automate test program generation for enhanced defect coverage, fewer test patterns, and faster time to results.
One of the inherent advantages of using AI for silicon chip designs is, as the tools learn from design to design, they get better. And that’s great news for the chip design world and for everyone who depends on electronic products. AI can empower engineers to get the right chip with the right specs to market faster—and to create more complex systems to tackle our world’s complex problems.
Gain more insights into AI-driven chip design and verification from these recent blog posts: