Episode 67: A Conversation with Aart de Geus on EDA and the Future of Silicon Design

The Circuit · 13 May 2024 · 42m · Episode 68

AI in practice: 3 real uses of AI described

What this episode covers

Aart de Geus, co-founder and executive chair of Synopsys, discusses the evolution of electronic design automation (EDA) and the future of silicon design. He explains how the industry shifted from analog to digital design, enabling the automation that drove Moore's Law, and details the transition to the Angstrom era with technologies like FinFETs and backside power delivery. De Geus also highlights the rise of systemic complexity through chiplets and vertical stacking, and the critical role of AI in optimizing these complex multi-die architectures.

AI in practice

Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.

DesignSynopsys AI toolswell over 300 production designs

Synopsys used AI pattern recognition to optimize chip design flows, resulting in designs that were substantially better and much faster than human-designed ones.

“And the first time that we saw real results was in late 2019 already, where we saw, hey, you know, we can do circuits substantially better and much faster than humans could. And today, some four years later or so, we have well over 300 production designs that are being done with that.”

Aart de Geus · Hear it at 13:47

DesignSynopsys AI toolstools got easily 20% better

Synopsys applied AI pattern recognition to optimize individual tool outputs, improving tool performance by approximately 20%.

“and the answer was definitely, yes, and the tools got easily 20% better during that time. But then came the next step. And that's the one that I'm really proud of, because synopsis has been absolutely an AI pioneer in that direction.”

Aart de Geus · Hear it at 13:47

DesignSynthesis (General Electric/Synopsys)30% smaller, and 30% faster

An early automated synthesis tool replaced manual design tasks, producing designs that were 30% smaller and 30% faster than human-engineered equivalents.

“And so we would go to customers and ask them for a design they were working on, and then pass it through the synthesizer, and literally, in a few hours, come back with something that was 30% smaller, and 30% faster. Power, at that time, was not a consideration.”

Aart de Geus · Hear it at 19:41

Mentioned

Companies
Synopsys, Intel, General Electric, ASML, Apple, NVIDIA, AMD
Tools
Spice

Other shows on the same ground, closest first.

  • EE Times Current · 19m

    The State of Analog Design in 2023

    Brett Brune, editor-in-chief of EE Times, and Majeed Ahmad, editor-in-chief of Planet Analog, discuss the 'Analog Everywhere' special report, analyzing the state of analog design in 2023. They explore the tension between integrating analog blocks into SoCs and the need for standalone devices in IoT and automotive sectors, while highlighting the evolution of EDA tools and the challenges of automating analog design. The conversation covers specific industry players like Analog Devices, Synopsys, and TSMC, and technologies such as FinFET and Gate-All-Around (GAA) processes.

  • Moore's Lobby · 55m

    Inside a Multi-$Billion IP Engine Driving IC and EDA Innovations

    John Koeter, Senior Vice President of IP at Synopsys, discusses the evolution of semiconductor intellectual property, the shift to multi-die chiplet architectures, and the transformative role of AI in chip design. He details Synopsys' acquisition of Ansys to enhance thermal modeling for 3D ICs, the challenges of managing heat in stacked dies, and the critical importance of software ecosystems for emerging architectures like RISC-V. Koeter also offers career advice on adaptability and leveraging AI as a force multiplier for engineers.

  • Moore's Lobby · 1h 01m

    Synopsys' Stelios Diamantidis Is Bringing the Magic of AI To Chip Design

    Stelios Diamantidis, Distinguished Architect at Synopsys, discusses the evolution of chip design from manual processes to AI-driven optimization. He recounts his career journey, including founding two EDA startups focused on verification and Silicon Lifecycle Management (SLM), before joining Synopsys to lead the development of DSO.ai. Diamantidis explains how reinforcement learning is used to solve intractable physical design problems, optimizing power, performance, and area (PPA) for diverse verticals like CPUs, GPUs, and automotive chips.

  • The Amp Hour · 1h 19m

    An Interview with Steve Leibson - Impavid Ideopraxist Insider

    Steve Leibson, marketing director at Cadence and author of the EDA360 Insider and Denali Memory Report, recounts his career from designing early HP computers like the HP 9825 and HP 9845 to working at early EDA firms like Cadnetics and later as a technical editor for EDN Magazine. He discusses the evolution of the semiconductor industry, the challenges of scaling to 20nm and below, the future of memory technologies like magnetic RAM and memristors, and the business dynamics of the EDA market. Leibson also shares anecdotes about the Design Automation Conference (DAC) and his unexpected appearance on a TV show with Leonard Nimoy.

  • EE Times Current · 40m

    An All Inclusive Journey to HPC/Data Center Silicon Success

    Charlie Matar, Senior Vice President of Systems Solutions at Synopsys, discusses the escalating complexity of High Performance Computing (HPC) and data center silicon design. He highlights critical challenges including engineer shortages, the high cost of advanced nodes, and the shift toward chiplet architectures and 2.5D/3D packaging. Matar also explains how Synopsys is leveraging its DSO.AI platform to optimize power, performance, and area (PPA) while addressing reliability and security concerns in the semiconductor ecosystem.

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Circuit. The episode itself belongs to its publisher.