Blog Review: Aug. 19

Countering sophisticated chip attacks; shifting left on auto SoCs; embedded OS power; computational storage.


Rambus’ Scott Best digs into some of the most sophisticated attacks used to target and compromise security chips, such as laser voltage probing, focused ion beam editing, reverse engineering, and NVM extraction, and ways to counter them.

Synopsys’ Chris Clark proposes a way to identify problems earlier and better ensure safety and reliability in automotive SoCs by moving from a linear development process to a parallel one.

Mentor’s Colin Walls argues that software developers have an important role in keeping the power consumption of embedded devices as close to the hardware’s minimum as possible and the two main factors to look at when choosing an embedded OS.

Cadence’s Paul McLellan listens in to a keynote by Nafea Bshara of AWS’ Annapurna Labs on the move to cloud-based EDA tools and how it’s changed team organization.

Arm’s Neil Werdmuller examines the rise of computational storage, which allows data to be processed directly on the drive where it is stored, and range of applications where it could lower latency, reduce energy use, and improve security.

SEMI’s Jim Hamajima checks out the impact COVID-19 has had on Japan’s semiconductor industry, which markets are appearing to expand or contract, and growing acceptance of work-from-home.

Ansys’ Gwenaël Moysan and Surrey NanoSystems’ Michael Stellmacher consider why the ultra-dark coating Vantablack may be useful to a project and how to simulate the impact of its use.

Nvidia’s Scott Martin profiles Fiddler Labs, a startup working on an explainable AI platform that can point to why decisions were made by an AI model based on the data inputs and their weighted values that were used to arrive at the data output.

Plus, don’t miss the blogs featured in last week’s Low Power-High Performance newsletter:

Editor in Chief Ed Sperling examines new technologies and approaches that could boost performance by orders of magnitude.

Synopsys’ Stelios Diamantidis contends that the design-by-optimization paradigm calls for a new way of looking at how hardware and software interact.

Fraunhofer EAS’ Christoph Sohrmann explains how standardization efforts are underway to ensure information can be accurately shared through the automotive supply chain.

Mentor’s Akshay Sarup and Mark Olen prescribe saving time by eliminating the need to code a new sequence for each scenario.

Rambus’ Frank Ferro concludes that continued increases in memory capacity and bandwidth are needed to keep AI accelerators and processors from being bottlenecked.

Ansys’ Krista Loeffler explains how to ensure that components of a safety system not only work as designed but also operate appropriately in real-world scenarios.

Cadence’s Tyler Lockman digs into the importance of automatically correcting sharp angles and acid traps with minimal change to routing layers.

Arm’s Rob Aitken focuses on the impact of data analytics and AI on global energy use and what can be done to improve it.

Moortec’s Richard McPartland warns that increased process variation and the end of Dennard scaling combine to mean the worst case for device power is getting worse.

Leave a Reply

(Note: This name will be displayed publicly)