Blog Review: May 17

Ransomware attack; FD-SOI; formal in China; brain-implantable SoCs; medical device security; CNN vectorization.

popularity

Synopsys’ Robert Vamosi digs into last Friday’s massive ransomware infection that impacted the UK health system, a Spanish telecom, and many other organizations running unpatched Windows – and whether there’s a second version out there.

Cadence’s Paul McLellan reports on the latest developments and future of FD-SOI from the SOI Silicon Valley Symposium.

Mentor’s Joe Hupcey III chats with colleague Jin Hou about DVCon China and the state of formal in the country.

A Lam Research author presents a primer on the atomic layer deposition process and a few of its applications.

ARM’s Karthik Ranjan announces a research project with the University of Washington to develop a brain-implantable SoC for bi-directional brain-computer interfaces aimed at solving neurodegenerative disorders.

Rambus’ Aharon Etengoff notes that the FDA is becoming increasingly concerned about security vulnerabilities in implantable medical devices and the dangers of side-channel attacks.

Intel’s Matthew Rosenquist argues that it’s time for the cybersecurity community to begin discussing the risks and opportunities of artificial intelligence, which could open a Pandora’s box for malicious attackers.

GlobalFoundries’ Mark Granger chats about the changes taking place in the automotive industry from a foundry perspective.

Samsung’s Kelvin Low points to what to look forward to at the upcoming Samsung Foundry Forum on Wednesday, May 24 at the Santa Clara Marriott.

Nvidia’s Samantha Zee points to a new application for AI: real-time video captioning of American Sign Language.

Silicon Labs’ Lance Looper notes that while switching technology using gallium nitride or silicon carbide allows for much faster switching rates, they require isolated gate drivers to work.

In a video, Cadence’s Megha Daga discusses ways to get the maximum compute power out of a convolutional neural network with different vectorization schemes.

The latest video on embedded software from Mentor’s Colin Walls discusses the C keyword static and its different, rather confusing, meanings.

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

Editor In Chief Ed Sperling examines how quickly the switch to autonomous vehicles can occur.

Executive Editor Ann Steffora Mutschler contends it is more important than ever to understand dynamic and leakage power and their place in the design flow.

Cadence’s Samer Hijazi pushes deeper into solving power limitations for convolutional neural networks on DSP processors.

Rambus’ Mohit Gupta finds Ethernet is moving faster than ever, presenting a distinct set of challenges for SerDes designers.

ARM’s Soshun Arai and Recognition Technologies’ Mark Sykes argue that making voice recognition work in the car requires the best of both local processing and the cloud.

Synopsys’ Ralph Grundler digs into adding flexibility to design by supporting multiple protocols in an interface system.

Mentor’s Mike Santarini looks at why some EDA tools have kept the top spot for more than a decade.

DAC general chair Michael MacNamara looks at what’s on tap at next month’s Design Automation Conference.