Blog Review: April 19

Portable stimulus and languages; from universities to commercialization; compromised emergency systems; beyond CNNs; cryogenic computing.

popularity

Mentor’s Tom Fitzpatrick explains what the Portable Stimulus standard will do, what it won’t, and why the choice of input language defined by the standard matters.

Cadence’s Paul McLellan listens in as IRDS chairman Paolo Gargini explains how long it takes technology breakthroughs to make out of the lab and into high-volume manufacturing.

Synopsys’ Robert Vamosi points to the recent sounding of all 100+ civil defense sirens in Dallas, Texas as just one reason to modernize legacy means of access.

Intel’s Ron Wilson looks beyond the standard convolutional neural network to examine other promising architectures for machine learning.

Rambus’ Aharon Etengoff takes a look at what cryogenic computing and very cold memory systems could mean for the performance and power demands of data centers.

Semico Research’s Morry Marshall contends that while 200mm fabs may use older technology, they’re far from passé.

A National Instruments staff writer looks at the layered architecture designed to make sure Industrial IoT assets have access to the right data at the right time.

ARM’s Brian Fuller checks out what technology is in the mountain climbing kit of ARM engineer Matt du Puy as he prepares to summit Kanchenjunga.

A Lam Research staff writer celebrates Earth Day by pointing out a few ways that new technologies are helping to drive energy efficiency.

In a video, Cadence’s Tom Hackett explains why the USB Type-C interface took longer than expected to show up in consumer electronics and five barriers that apply to the adoption of new standards.

Mentor’s Andrew Pattersen argues that V2X technology supporting high-priority road-traffic information will become a standard production feature of many vehicles in the next decade.

And don’t forget the blogs highlighted in last week’s Low Power-High Performance newsletter:

Editor In Chief Ed Sperling polls four CEOs on why security adoption is so slow, and questions the impact on power and performance.

Executive Editor Ann Steffora Mutschler observes that when it comes to understanding electricity, sometimes a down-to-earth explanation is best.

Cadence’s Samer Hijazi digs into the limitations of convolutional neural nets on DSP processors.

Rambus’ Bill Fuller finds managing signal integrity increasingly difficult as data rates move to 56Gps and beyond.

ANSYS’ Ravi Ravikumar contends that automotive ICs need to be verified and validated in the context of the entire system, over billions of miles of road testing.

Mentor’s Progyna Khondkar examines how to handle design-specific power-aware verification complexities with SystemVerilog and UPF.

ARM’s Govind Wathan explains that moving AI out of the cloud requires a new approach to mobile architectures.

Synopsys’ Angela Raucher argues that the long lifecycle of automotive ICs means keeping functional safety requirements in mind.