What Engineers Are Reading And Watching

By Brian Bailey And Ed Sperling
An important indicator of where the chip industry is heading is what engineers are reading and what videos they are watching. While some subjects remain on top, such as the level of interest in the latest manufacturing technologies, other areas come and go.

The stories with the biggest traffic numbers are almost identical to last year. Readers want to know what is happening at 5nm and 3nm, and that by necessity involves EUV. The other category that garnered the most views involves memory. But there also is growing interest in power and chip architectures, as well as the increasing role of semiconductors in markets such as automotive and AI. And there is a rising concern about the reliability and interaction of chips and systems within those markets.

Manufacturing
The article with the highest readership in this category: 5nm Vs. 3nm — Foundry vendors are readying the next wave of advanced processes. Their customers will face a myriad of confusing options—including whether to develop chips at 5nm, wait until 3nm, or opt for something in between.

A potential split in the supply chain and the rise of continues to draw readers: China’s Foundry Biz Takes Big Leap Forward — 30 facilities are planned, including 10/7nm processes, but a number of factors could slow progress.

Also of interest is what’s happening in screen technology. MicroLEDs: The Next Revolution In Displays?  — Technology offers improved brightness, colors, and lower power, but they’re expensive and difficult to manufacture.

System-Level Design

This category went through an explosion in traffic numbers this year. While manufacturing numbers usually tend to be far higher than the other categories, some SLD articles gave them a run for their money in 2019.

The top traffic draw: The Case For Embedded FPGAs Strengthens And Widens — Combining the flexibility of a FPGA with the performance and cost benefits of an SoC is pushing this technology well into the mainstream.

System architectures, in general, were a hot topic: Chiplets, Faster Interconnects, More Efficiency — Why Intel, AMD, Arm and IBVM are focusing on architectures, microarchitectures, and functional changes.

FPGAs and embedded FPGA were a part of that trend, as well, in part because there is so much churn in new technologies: Using FPGAs For AI — How good are standard FPGAs for AI purposes, and how different will dedicated FPGA-based devices be from them?

Interest in RISC-V continued to grow this year. The Challenge Of RISC-V Compliance — Showing that a processor core adheres to a specification becomes more difficult when the specification is extensible.

This also has ignited interest in the possibilities for open source EDA: Will Open-Source EDA Work? — DARPA program pushes for cheaper and simpler tools, but it may not be so easy.

A new and popular category for this year is in- and near-memory computing: Using Memory Differently To Boost Speed — Getting data in and out of memory faster is adding some unexpected challenges. The topic drove high readership in other slices of this approach: In-Memory Computing Challenges Come Into Focus, Will In-Memory Processing Work?, and Solving The Memory Bottleneck. There was a popular video, as well: In-Memory Computing.

Moving large amounts of data around a system is no longer the path to success. That approach is too slow and consumes too much power. It may be time to flip the equation.

Packaging and power
In the past, packaging and power were only marginally connected. Much has changed since 28nm, and increasingly at every node since then. The big driver for advanced packaging is the inability to obtain performance and power benefits from scaling alone.

This is why advanced packaging, 3D design and chiplets are becoming important topics. Notable articles include HBM2E: The E Stands For Evolutionary — The new version of the high bandwidth memory standard promises greater speeds and feeds, particularly in 2.5D and fan-out configurations. Another popular story on this issue over the past year: Chiplet Momentum Builds, Despite Tradeoffs — Pre-characterized tiles can move Moore’s Law forward, but it’s not as easy as it looks.

Alongside of this, power emerged as the gating factor in advanced chip designs. It’s also getting more difficult to manage, as evidenced by Power Complexity On The Rise — New architectures, different markets and more variables make it increasingly difficult to design and verify low-power chips.

Some popular videos in this space: Multi-Physics at 5/3nm, GDDR6 — HBM2 Tradeoffs.

Automotive, security, pervasive computing
The electrification of vehicles has turned a niche industry in automotive electronics into a global boom. It also has pushed semiconductors firmly into the world of functional safety, where reliability is a key issue. Chips have to continue to function in extreme conditions, and they have to function to spec for a decade or more.

Data is key in this world, as evidenced by the strong readership in Using Sensor Data To Improve Yield And Uptime and Why Analog Designs Fail.

It also has raised a number of questions about what’s necessary to make chips that can endure extreme operating conditions over time: How To Build An Automotive Chip — Changing standards, stringent requirements and a mix of expertise make this a tough market to crack. This is a complex market with unique demands, such as the weight of the electronics, (Shedding Pounds In Automotive Electronics) and how to deal with various types of noise (Why Chips Are Getting Noisier), which became popular topics for readers in 2019.

Alongside of this are other burgeoning markets, such as 5G, IoT, and the edge, which fits between the cloud and the end point. Cars are edge devices, but the edge encompasses much more than automobiles. So far, though, the definition of what fits into the is somewhat vague, and that has generated a lot of questions and strategies, as well as strong readership for stories such as, AI Chip Architectures Race To The Edge, and Revving Up For Edge Computing.

5G has created a slew of questions, as well, particularly in the millimeter wave space. This area is still evolving, as seen in Issues In Designing 5G Beamforming Antennas, and 5G Heats Up Base Stations — Inefficient conversion of RF to digital and continuous connectivity issues are causing thermal problems, threatening signal integrity and reliability.

Popular videos across these subjects: Safety Critical Design In Automotive, Using Multiple Inferencing Chips In Neural Networks, Improving Quality Through Data Analytics, Which Verification Engine When, and Signoff-Compatible CDC.

Less obvious, but a strong traffic driver in 2019, is how security intersects with reliability. Traffic was strong in 2019 on security-related issues, such as Finding Hardware Trojans.

Other popular videos: Advanced Process Control; Meltdown, Spectre and Foreshadow; and Machine Learning Inferencing At The Edge.

Future tech
What’s on the drawing board is always a strong traffic driver, as well: Get Ready For Integrated Silicon Photonics — This More than Moore technology is still ramping up, and problems need to be solved, but it could lead to some fundamental changes.

And finally, well into the future, the semiconductor industry is preparing for the quantum era: U.S. Consortium Pulls Ecosystem Into Quantum — Through QED-C, the U.S. government helps build industry support behind quantum computing.

Stay tuned for much more in 2020. Change is accelerating, and understanding how the pieces fit together will be a continual problem throughout the next decade.