Synopsys: 2021 Predictions for EDA Innovation

As the leading supplier of electronic design automation (EDA) solutions, Synopsys provides chip designers with solutions and methodologies to address the dynamic challenges of hardware design. Looking at the year ahead, we see some trends in a few key application areas that will be interesting to track.

Artificial Intelligence Calls for Innovative Design Solutions

With the proliferation of artificial intelligence (AI) applications comes a need for more innovative hardware architectures. While these architectures are unprecedentedly complex and large, performance-per-watt remains critical for efficient operations. For this, our AI experts, Arun Venkatachar and Stelios Diamantidis, anticipate broader adoption of enhanced versions of existing tools. It’s time for existing tools to become more architecture-aware and feature new capabilities and approaches that dramatically accelerate implementation of new compute paradigms. For example, consider design implementation solutions that connect with the designer’s requirements at a higher level of abstraction to allow an accurate estimate of power, performance, and area (PPA) during the architectural exploration phase. Such solutions would be most beneficial if they follow a highly convergent path through to final signoff for manufacturing.

Also interesting is how AI is impacting the chip design process itself. For example, last year, Synopsys introduced DSO.ai (Design Space Optimization AI), the world’s first autonomous AI application for chip design. With DSO.ai, chip designers can benefit from enhanced productivity while producing better chips via a new, generative optimization paradigm that uses reinforcement learning technology to autonomously search design spaces for optimal solutions.

Software Drives Automotive Applications

Our automotive experts Chris Clark and Dennis Kengo Oka foresee that a key focus in the automotive world this year will be on functional safety and security standards, such as ISO 26262 Road vehicles – Functional safety and ISO/SAE 21434, which addresses cybersecurity in vehicles. EDA and IP solutions that are certified for these standards can help meet quality requirements and accelerate the development of safety-critical automotive applications.

Clark and Kengo Oka also anticipate a continued increase in the use of software in vehicle designs. Automakers are now consolidating smaller microcontrollers into larger systems, running multiple operating systems and much more software content. The industry could see more OEMs and suppliers establishing their own software development centers or acquiring or partnering with other software companies. Carmakers may also continue leveraging methodologies that are common in the IT world, such as agile development, as well as developing new services and solutions based on web apps, mobile apps, and cloud platforms that interact with vehicles.

High-Performance Computing and Cloud

High-performance computing (HPC) and cloud computing technologies continue to power some of the latest innovations in areas like AI, autonomous driving, 3D printing, and facial recognition. However, with more people working, learning, and getting entertainment from home during the pandemic, these technologies have become increasingly important for everyday consumers. Synopsys thought leaders have some interesting perspectives on what might be happening in the HPC and cloud industries in 2021.

For Scott Durrant from Synopsys’ Solutions Group, reducing latency will continue to be critical. He explains: “There’s already been a lot of development around addressing latency, including the sheer processing power that’s been incorporated into compute devices to reduce processing time. But other technologies, including 400 Gigabit per second and faster Ethernet, allow us to move data much more quickly as well.”

Scott Knowlton, also from the company’s Solutions Group, agrees that system performance and getting data through the system faster are important trends, noting: “It’s driving server chips, it’s driving an increase in clock speeds, and it’s pushing to smaller geometry technologies at a pretty rapid rate—all of which makes the designs more complex. We’re also seeing an aggressive push in a number of technologies, such as the die-to-die interconnects in addition to chip-to-chip interconnects like PCIe and CXL, all being driven very quickly to move to the latest generations of protocols.”

More compute power will be pushed to the edge, where actual data is being received, said Ruben Molina from Synopsys’ Digital Design Group. “For example, in autonomous driving, the car is taking in a lot of data and has to make decisions very quickly, so it can’t wait for the information to be sent to a compute server; instead, it needs to be processed at the edge. We will also see increased edge computing at places like manufacturing facilities, where the health of production line automation is continually monitored to keep production lines from going down unexpectedly,” he said.

Susheel Tadikonda from Synopsys’ Verification Group offers his take: “Verification is not just going to be at an SoC level; system-level verification will become very important, especially as you look at the multi-dies and interconnect fabrics. It is very difficult to verify thousands of cores that come together on a single die and across dies. When you have a device that has two different chips from different generations, it is a lot more challenging to verify. That’s one of the things Synopsys is working on along with chip manufacturers to help solve these verification challenges.”

Now, as the months unfold, it’ll be worthwhile to note how these predictions will play out.

Tags: Predictions

This entry was posted on Tuesday, January 19th, 2021 at 5:55 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.