Reflections On 2017: Manufacturing And Markets

Part One: How close were the predictions for 2017? Most were fairly close but some big outliers and some things that were missed.


People love to make predictions, and most of the time they have it easy, but at Semiconductor Engineering, we ask them to look back on the predictions they make each year and to assess how close to the mark they were. To see what they missed and what surprised them. Not everyone accepts our offer to grade themselves, but many have this year. This is the first of two parts that looks at the predictions associated with semiconductor manufacturing and end markets. The second part will cover tools and methodologies.

Semiconductor manufacturing
To get things rolling, Aki Fujimura, CEO of D2S stated that Moore’s Law would most likely continue unabated for the next 10 years. While this comes as no surprise, what has surprised a number of people is the rate at which adoption of the latest nodes is happening. The economic benefit of following Moore’s Law is no longer evident, so generations of successive feature shrinks have given way to fundamental rethinking of the best way to architect a system. The first indication that change is underway is the increasing heterogeneity in systems.

Fujimura predicted changes on the mask side of production. He expected a transition from variable-shaped beam (VSB) technology to multi-beam technology. “I don’t think there will be production masks out, possibly one or two, but the ecosystem will be getting ready for it.” Indeed, it turns out that “multi-beam mask writing ecosystem have not yet been announced in production. However, based on the annual eBeam Initiative Perceptions Survey, which polls opinions from industry luminaries from across the semiconductor ecosystem, multi-beam mask writing is definitely about to happen.”

The reason is that “the overwhelming majority of masks that mask shops have to write today are still better off being written with variable-shaped beam (VSB) machines because the shot counts that are needed to write the mask are low enough, and the resist needed to expose the smallest shapes on the mask are fast enough for VSB machines. However, this will change with newer generations of leading-edge complex masks.”

Wally Rhines, CEO of Mentor Graphics chose to focus on packaging and the need for better tools. “Ultimately, the next generation of complexity will require simulation, verification and analysis of multi-chip packaging configurations.” Rhines noted that progress has been made.

However, in a recent article, Herb Reiter, president of EDA2ASIC Consulting noted that “as we are seeing more of the value creation shifting to the package and the cooperation between package and silicon, we are running into a big headache because there is no such thing yet as a die-package co-design flow. It’s very difficult to feed information from the silicon world into the packaging world, and even more difficult to get feedback from the packaging world information into the silicon world to co-optimize these two domains.”

There is a lot of development going on in this area, with more expected in the future. As Rob Knoth, product management director for Cadence‘s Digital & Signoff group, notes “the role that engineering innovation and semiconductors play in making the world a better place doesn’t get a holiday or have an expiration date.”

Fujimura adds some other observations about 2017: “The semiconductor industry is getting ready to use EUV to expose production wafers. Initial deployment will be in places where multiple patterning with 193nm is already deployed for production. This allows the eco-structure to make a smoother transition instead of requiring a sudden step-function transition of everything at once. For example, in the initial deployment of EUV a mask pellicle may not yet exist and actinic mask inspection (inspecting EUV masks using an EUV wavelength source) will not be ready.”

One of the big concerns voiced by many people is that security, or the lack of it, could threaten the advancement and adoption in each of the growth areas for semiconductors, including Internet of Things (IoT), automotive and artificial intelligence. “There is a lack of serious effort and understanding to implement security in hardware except in the metering, financial, and mobile markets where standards are well-defined,” said Ron Lowman, strategic marketing manager for IoT at Synopsys. “Most developers have identified that true random number generators are required in any security implementation. Many standards organizations are also becoming more aware that software alone will not be enough. This is evident by additional implementations of unique ID, multi-factor authentication and more. Many architects are now looking much more seriously at their products threat levels to define their security implementations from the ground up.”

Interestingly, progress in security could have an upside. “Security will create growth, driven by the need to contain random, politically motivated attacks on people and infrastructure,” says K. Charles Janac President & CEO for ArterisIP. “The non-believers have to be protected from the fanatics. The ‘haves’ must be shielded from the ‘have-nots’. Security will drive deployment of camera networks, very large-scale data centers, big data deep learning analysis and anti-hacking measures. Accelerators, image processing, and deep learning will drive numerous semiconductor applications in this space. I am afraid that we will be trading security for privacy – powered by silicon and software. Overall, the semiconductor industry will do well as long it is nimble and global enough to adjust to geographic and application rotation.”

One area that had been suffering was automotive, which had seen some high profile security lapses. That suggested to David Kelf, vice president of marketing for OneSpin Solutions, that automotive designs would need to be built to be as secure as possible and that tools to help predict vulnerabilities would become more important. “While I haven’t observed any big stories this year on vehicles being sent off the road due to malicious action, it is clear the concern is still there,” says Kelf. “At various workshops and other events, the subject of automotive security and security in other applications comes up often. The ISO 26262 committee continues to discuss how this should be included in the regulations, and for sure we will see new directives on this. Security is also an area of interest in the design and verification solutions space. Other security initiatives will become apparent in 2018 also in various applications.”

The IoT has seen some interesting changes this year, as predicted by Frank Schirrmeister, senior product management group director in the System & Verification Group at Cadence. He predicted late last year that “value shifts will accelerate from pure semiconductor value to systemic value in IoT applications. The edge-node sensor itself may not contribute to profits greatly. But the systemic value of combining the edge node with a hub accumulating data, and sending it through networks to cloud servers in which machine learning and big data analysis happens, allows for cross-monetization”

Schirrmeister says this is already happening. “The shift from pure semiconductor value to systemic value in IoT applications has certainly come to fruition and will most certainly advance further in 2018. In a recent blog, I outlined that the ‘mindset’ of developers is still one of the five key potential pitfalls for IoT proliferation. Users are capturing systemic value by applying machine learning to the collected dataset to optimize IoT usage even further.”

Another prediction related to integration within IoT. “There is a clear trend away from using separate wireless combo chips toward integrating connectivity onto the main application processor as companies look to reduce power, silicon area, and bill-of-materials costs for connected devices,” Jen Bernier, director of technology communications for Imagination Technologies, predicted last year.

Vikram Gupta, senior vice president at Cypress, believes it may be more complicated. “As the IoT market continues on an accelerated path, the industry faces a major challenge—developing right-sized chip architectures for the IoT. Rather than taking solutions that were developed for other market segments and repackaging them for IoT, a better approach is to develop architectures that are purpose-built for IoT. This means offering low-power MCUs with built-in security and leading edge wireless connectivity, delivering optimal solutions for the customer.”

Synopsys’ Lowman had foreseen problems caused by communications standards. “Many smart home and wearable applications have two issues. First, standard connectivity options are not interoperable enough, and Bluetooth 5, NB-IoT, and the coming of 5G will help remedy this, and second they don’t provide enough value to users beyond connectivity.” He responds, saying “Bluetooth 5 adoption continues to expand its application space beyond hearables and nearables to the home, office, and factory. 5G, LPWAN, and other wireless integration was, and still is, a hot topic. However, for the massive number of small devices there is still a huge diversity of interoperable solutions being aggressively positioned by different regions, different applications, different operating providers, different spectrums, and different suppliers causing confusion among developers, let alone the choices for consumers. Wireless integration adoption sees some hurdles due to perceived risks, perceived costs, and perceived maturity of technologies but this should clearly reverse course in 2018. I expect wireless integration and the components needed to still dominate design discussions in 2018 but with more expertise devoted to the value and implementations. I also believe the introduction of new Combo Wireless solutions will tie new and old, that can consolidate green field implementations moving forward to a smaller set of wireless options.

Schirrmeister also predicted that 2017 would be an interesting year for processor architectures. At that point he had said that “even Open Source hardware architectures are looking like they will be very relevant. It’s definitely one of the most entertaining spaces to watch in 2017 and for years to come.” He now responds. “The battle is raging and has resulted in some interesting new business models. MIPS has been spun out as part of the overall Imagination transition, and RISC-V certainly enjoys A LOT of attention with more companies endorsing it and announcing serious product developments for it. This trend will only accelerate in 2018 and has forced some of the traditional players to react fast—like Arm did with its Design Start initiative.

“The semiconductor industry witnessed a consolidation slowdown with new startups offering free, open solutions for today’s design challenges – not to mention established companies moving away from closed architectures,” adds Rick O’Connor, executive director of the RISC-V Foundation. “There is a growing interest in open-source instruction set architectures (ISAs), such as RISC-V. The portability and flexibility of the RISC-V architecture has driven innovation in a number of applications, addressing the increasing demands of our connected world from big data to the IoT. This newfound freedom in silicon design has also encouraged collaboration across the ecosystem by fostering a system-level approach to SoC design.

Even within traditional markets, processors are changing. “Now more than a decade into the modern smartphone era, we’re seeing mid-range devices catching up to flagship premium smartphones in terms of power and features,” says Finbarr Moynihan, General Manager, MediaTek. “OEMs are taking advantage of affordable, versatile components to design powerful smartphones that have the big, new features consumers are looking for, without the premium price tag. In 2018, more BIG cores will be hitting affordably priced, many-core smartphones delivering remarkable performance, huge gains in energy efficiency and unprecedented value.”
What they missed

One development that everyone missed is the increase in local processing. “While traditional mobile AI applications have performed all or most processing in the cloud, powerful chips are now able to handle compute-intensive processes right on consumers’ devices,” adds Moynihan. “Energy-efficient chipset designs have been critical in making this a reality, enabling local processing without a huge drain on battery life. Local neural net data processing is poised to help revolutionize the smartphone experience as OS-level processes are optimized through machine learning and smartphones predict what consumers want to do, delivering a more refined experience than ever before.”

Lowman adds “Voice Recognition, and the associated processing required, clearly continues adoption in the market and many times can replace displays as an HMI (human machine interface). The minimum processing requirements are consistently being reviewed to understand future architecture designs and power budgets.”

Lowman points to one area that he had hoped would see developments in 2017, but failed to meet his expectations. “Energy Harvesting MCUs didn’t make a major impact on applications, yet, other than some small wearables but this space will likely continue to grow as the next generation of systems is developed with these solutions that includes the proper power management and other system components needed such as wireless.”

Nobody was prepared to predict what would happen in the consolidation space at the beginning of 2017. Arteris’ Janac tries to summarize the events of the year. “2017 was an exciting transition year for geographic and application rotation. Semiconductor mergers in USA, Europe, and Japan continued because of slowing market growth: If you can’t grow internally, you buy. In contrast, there were multiple well-funded semiconductor start-ups created in China.”

Semiconductor companies in China remind Janac of the optimistic spirit of Silicon Valley of the 1980s: “they have the enthusiasm to allow them to overcome inexperience. They work hard and move fast. Proven, production-ready IP will help their efforts. Creative semiconductor destruction in China will enable the best-of-breed to recycle unsuccessful assets to become global players, and it will make the country a source of growth over the next few years. Companies that successfully align with this trend will prosper while those that do not may suffer.”