Week In Review: Design, Low Power

Renesas’ new acquisition; NVIDIA partnerships; EDA’s big changes; battery drain tool; complexity of 3D designs; chiplets integration & mini-consortia; China build out; thermal management; quantum computer deployment.

popularity

Renesas will acquire Panthronics, a fabless semiconductor company specializing in high-performance wireless products, expanding its reach into near-field communications for financial, IoT, asset tracking, wireless charging, and automotive applications. The two companies already had collaborated on designs for mobile point-of-sale terminals, wireless charging, and smart metering.

Renesas also rolled out an industrial MPU for servo drives, inverters, and industrial and collaborative robots. Target markets include factory and building automation, and medical equipment.

Keysight introduced a battery emulator and profiler that helps identify the impact of variables –such as hardware changes, software updates and temperature changes– affecting battery drain on IoT devices.

MobileDrive, a joint venture between FIH Mobile and Stellantis, has adopted Siemens Xcelerator portfolio and its digital twin technology to supports its ADAS technology roadmap.

IBM deployed a quantum computer at the Cleveland Clinic, claiming it’s the world’s first quantum computer dedicated to healthcare research onsite in the private sector.

At its GPU Technology Conference (GTC) this week, NVIDIA announced multiple partnerships to bring AI, simulation and other collaboration capabilities to multiple industries. Among the collaborations:

  • Synopsys, ASML, and TSMC are integrating NVIDIA’s cuLitho software library for computational lithography into their respective processes and systems. Synopsys will run its optical proximity correction (OPC) software on the cuLitho software library. “Running on GPUs, cuLitho delivers a performance leap of up to 40X beyond current lithography — the process of creating patterns on a silicon wafer — accelerating the massive computational workloads that currently consume tens of billions of CPU hours every year,” NVIDIA said.
  • Oracle will host NVIDIA’s new DGX Cloud infrastructure. DGX cloud provides dedicated clusters of DGX AI supercomputing. Microsoft and Google Cloud are expected to begin hosting DGX cloud later this year, and
  • NVIDIA and Quantum Machines are building a new system for accelerating quantum-classical computing.

Evolving EDA

With rapid changes in chip architecture, end markets, geopolitical tensions and new technologies, the EDA industry is being constantly challenged to adapt. Here’s a look at how the industry is meeting the demands.

Case in point: creating real 3D designs is proving to be much more complex and difficult than 2.5D, requiring significant innovation in both technology and EDA and other tools.

Chiplets

Creating chiplets with as much flexibility as possible has captured the imagination of the semiconductor ecosystem, but how heterogeneous integration of chiplets from different foundries will play out remains unclear.

Mini-consortia for chiplets are sprouting up across the industry, driven by demands for increasing customization in tight market windows and fueled by combinations of hardened IP that have been proven in silicon.

China

China startups 2022 Semiconductor Engineering E-Book CoverChina is racing to stay competitive in semiconductors, as trade sanctions increasingly limit its access to EDA tools and manufacturing equipment required for the most advanced manufacturing processes. As a result, state-backed investors and regional development funds are pouring money into domestic semiconductor companies.

China-based RISC-V company StarFive received a strategic investment from Baidu, bringing the total raised amount from all financing to $146 million. StarFive’s RISC-V products will be integrated in Baidu’s data centers.

Marvell eliminated its R&D team in China, as part of its latest round of job cuts, according to the South China Morning Post.

Energy and Heat

DARPA is looking for new approaches and technologies for dynamic thermal management to address various application conditions and usage environments. Responses are due by March 31, 2023.

With the cost of energy rising this past year and data centers energy consumption forecast to reach 8% of the world’s electricity by 2030, power consumption in data centers is getting much more attention. The University of Buffalo’s recent study showed data centers “can reduce energy consumption 10% to 30% by consolidating jobs to as few servers as possible when workloads are high, and evenly distributing the workload across all servers when loads are low.”

This data center challenge has led to a range of approaches to cooling. While no one can claim the problem is solved, there are refinements to older approaches, along with a few novel ideas that give some hope of better balancing demands.

On a smaller scale, by harnessing the heat of data centers, swimming pool centers are reducing their energy bills. The concept is relatively simple — hot oil is pumped into a heat exchanger to heat up the water in the pool.

Further Reading

Semiconductor Engineering just published the latest March Systems & Design newsletter. Find the full lineup here.

The March Low Power-High Performance newsletter includes these features and more:

  • Uneven Circuit Aging Becoming A Bigger Problem
  • MIPI’s Focus Widens
  • What Makes RISC-V Verification Unique

Events

Upcoming chip industry events can be found here and include:

  • International Symposium on Physical Design (ISPD), Mar. 26 – 29
  • MEMCon 2023, Mar. 28 – 29 in Mountain View, CA
  • SNUG Silicon Valley, Mar. 29 – 30 in Santa Clara, CA
  • DATE 2023, Apr 17 –19 in Antwerp, Belgium
  • Impact of New Regulations on the Semiconductor Design Ecosystem, Apr. 26 in San Jose, CA

 



Leave a Reply


(Note: This name will be displayed publicly)