Cloud native EDA tools & pre-optimized hardware platforms
For decades, Moore’s law enabled semiconductor companies to double the number of transistors in an integrated circuit (IC) every two years. As Moore’s law slows, device scaling for monolithic systems-on-chip (SoCs) is noticeably less while the cost of newer, more complex process nodes continues to steadily increase. However, the demand for faster, better, and smarter silicon is only growing as the exploding impact of “Smart Everything” and ubiquitous artificial intelligence (AI) drive a massive push for faster speeds and ever-more transistors. We call this intersection of Moore’s law scale complexity with new systemic complexity demands the SysMoore era.
Aart de Geus, 草榴社区 chair and chief executive officer, recently observed that semiconductor companies are revolutionizing both architectural function and form with new multi-die systems to help meet market demand. Indeed, disaggregating a large monolithic chip design into smaller, proven dies results in higher yields, lower silicon costs over time, and more customizable SKUs. de Geus, who penned the opening chapter of a new 草榴社区 industry insight report titled “How Quickly Will Multi-Die Systems Change Semiconductor Design?” notes that cross-industry collaboration has seen chip designers significantly increase connectivity density while reducing energy switching (per bit). The report also includes insights on multi-die systems from Ansys, Arm, Bosch, Google, Intel, and Samsung.
“As angstrom-sized transistors intersect with multi-die Si-substrates, we see classic Moore pass the baton to SysMoore,” writes de Geus. “Today, 草榴社区 tracks more than 100 multi-die system designs. Be it through hardware/software digital twins, multi-die connectivity IP, or AI-driven chip design, we collaborate closely with the leading SysMoore companies of tomorrow.”
In this blog post, we’ll review key takeaways from the new semiconductor industry insight report and summarize perspectives from leading corporations, with commentary from their senior executives, including:
As traditional Moore’s law scaling approaches its physical limits, Ansys sees the quest for higher electronic system densities transitioning to multi-die systems with the help of 2.5D and 3D packaging. However, the company emphasizes that successful adoption of multi-die systems hinges on overcoming three challenges: multi-scale, multi-physics, and multi-organizational coordination.
“Advanced multi-die systems condense three design scales into one design challenge that crosses six orders of magnitude, from nanometer IC design, through millimeter package design, to centimeter 3D-IC systems,” states Ansys. “These solutions are divided into three tool suites (IC, systems, and packaging) that need to be integrated into a single solution.”
Ansys says the potential for more capable multi-die systems with higher yields is already positively impacting suppliers of high-performance computing (HPC) processors and graphics, as well as AI and machine learning (ML) enablement at the cloud edge.
Arm expresses similar sentiments, noting that multi-die systems composed of co-packaged chiplets will become widespread across the industry. “Companies will be able to amortize their hardware and software engineering investment by reusing chiplets for multiple products,” the company says. “Complex systems will be cleanly partitioned, reducing risk, cost, and time to market.”
According to Samsung Electronics, both 2.5D and 3D packaging technologies such as Samsung’s I-Cube 2.xD and X-Cube 3D IC empower device manufacturers to pursue new product designs based on multi-die systems. “Breakthroughs in AI, 5G, autonomous vehicles, and metaverse tech promise to reshape the way we live—but delivering the function and performance needed to power those advancements on a single chip is becoming more complex and less cost-effective,” explains Samsung. “Combining the power and diversity of today’s chips in a unified system brings new products within reach.”
Intel notes that advanced packaging is already playing a crucial role in enabling multi-die systems. “Heterogeneous integration using advanced 2.5 and 3D packaging technologies, like EMIB and Foveros, offers an attractive path to architecting products by combining chiplets from various sources, designed on disparate silicon nodes,” the company states. “This mix-and-match capability unlocks the ability to optimize for unique functionality, performance, and cost while enabling reuse and modularity.”
Read our comprehensive eBook and learn how to overcome multi-die design challenges and realize success.
To maximize the impact of advanced packaging solutions and chiplets, Google says the semiconductor industry must look beyond traditional logic design and embrace modularity. “Chiplets allow us to do co-design in a multi-die system context, allowing cost advantages but also mix-and-match integration across heterogeneous IP blocks.”
Google also proposes a new framework to deliver advanced computing and to rewrite the rules of hardware innovation by “creating harmony” between hardware and software. “System-level optimization or co-design where we look at the whole stack—from the application level all the way down to the chip level—can achieve dramatic benefits.”
Intel notes that standardization is an essential prerequisite to creating an open, streamlined chiplet ecosystem. “Universal Chiplet Interconnect Express (UCIe) is one such standard—and a crucial step in enabling [a] rich and robust industry ecosystem.”
Bosch sees an open chiplet ecosystem as crucial for the automotive industry. The company explains that custom chiplet designs for vehicles can be based on proven silicon architectures for non-automotive, high-performance applications. “This will allow the break-up of a monolithic system into separate integrated circuits on multiple dies, integrated by advanced packaging technologies to comprise a high-performance compute unit.”
Bosch emphasizes that clearly defined interfaces, connections, and standards are essential for an open automotive chiplet ecosystem—as is support from a multitude of industry players that produce similar chiplets. “This approach delivers the highest degree of modularization and co-creation to enable best, scalable products.”
草榴社区’ Sassine Ghazi concurs, noting that streamlining and optimizing heterogeneous multi-die systems is an initiative that requires cross-industry collaboration. “Less effort equals lower cost,” says Ghazi.
Ghazi sees a flood of high-end multi-die systems spanning a wide range of markets by the 2030s. “I say this because of the rapid progress I see, not only in 草榴社区’ engineering labs but also the billions of dollars of secure investment going into multi-die systems from across public and private sources.”
Specific examples of multi-die system innovations include fusing architectural analysis and implementation as part of multi-die system flows, as well as the addition of AI for a step-function increase in productivity and quality of results. “Even now, our customers and partners are designing products that appeared completely out of reach just a few years ago,” says Ghazi.
As Ghazi points out, more than 95% of advanced chips today are built using 草榴社区 technology. “We’re tracking over 100 designs today, with an uptick of 20% over the last six months,” adds Ghazi. “That tells me this is a design direction which is maturing very fast indeed.”
草榴社区 is committed to working with ecosystem partners as we usher in an innovative era of multi-die systems. We can’t wait to see what the semiconductor industry accomplishes next with our AI-driven EDA tools, IP products, and deep system design expertise.
Get an industry-wide look at the accelerated adoption of multi-die systems. Access your copy of the new 草榴社区 industry insight report here.
草榴社区 is empowering technology visionaries with a comprehensive and scalable multi-die solution for fast heterogeneous integration.
Download Brief