Blog Review: Nov. 13

Co-design for AI; emerging NVM; optimizing aircraft; password limitations; pressure sensor’s origins.


Applied Materials’ Buvna Ayyagari-Sangamalli argues that the siloed structure that produced the computing eras of the past will not be sufficient to fuel the AI era and that a new codesign approach to everything from architecture to materials is needed.

Arm’s Wendy Elsasser examines emerging non-volatile memories and how they have triggered innovation for new memory protocols and optimized storage solutions, redefining the memory sub-system.

Mentor’s Sumit Vishwakarma traces the evolution of sensors from a 1923 tire pressure monitor to the key requirements of today’s MEMS sensors

Cadence’s Paul McLellan shares a talk on the intersection of formal verification and machine learning from the recent Jasper User Group meeting.

Synopsys’ Taylor Armerding argues that passwords are obsolete and security should turn towards alternatives such as biometric authentication or activating a physical token, but even where passwords are used improvements are possible.

ANSYS’ Paolo Colombo finds that small optimizations could have a big effect on the environmental impact of aircraft, with just 1% changes to weight, propulsion efficiency, and aerodynamics saving emissions and money, plus the importance of reducing aircraft noise.

SEMI’s Serena Brischetto chats with of Markus Hörburger of Atotech about improving the reliability of advanced heterogeneous packaging through innovative chemistry solutions for fan-out wafer-level packaging in 5G applications.

And don’t miss the featured blogs from the recent IoT, Security & Automotive and Test, Measurement & Analytics newsletters:

Editor In Chief Ed Sperling contends that while vehicles are getting smarter, full autonomy under any circumstances is still a long way off.

Arteris IP’s Kurt Shuler expounds on the virtues of creating a reliable place to manage critical functions when a design contains a mix of ASILs.

Synopsys’ Shivakumar Chonnad, Vladimir Litovtchenko, and Rohit Bhardwaj make the case for a systematic approach to requirements management, which can mean fewer project iterations and more efficient products.

Mentor’s Richard Pugh wants more focus on pre-silicon verification for automotive chips.

Flex Logix’s Geoff Tate looks at why TOPS may correlate with cost, but not necessarily with throughput.

Cadence’s Paul McLellan reminds us that sometimes much-hyped, apparently innovative avenues can lead nowhere.

Rambus’ Gary Kenworthy explains that no matter how costly or inexpensive, all physical electronic systems routinely leak information about the internal process of computing.

Editor In Chief Ed Sperling argues that how AI/ML systems will age is not well understood, and that’s a big problem.

Mentor’s Matthew Knowles describes erasing the gap between automatic test equipment and DFT debug software to streamline silicon bring-up and debug.

National Instruments’ Sarah Yost admits it’s exciting to look ahead to the next big thing but warns a lot is left to do in 5G.

OptimalPlus’ Peter Hodgins explains why data helps optimize the manufacturing process of electric powertrains to make them higher quality, safer, and less expensive.

YieldHub’s Kevin Robinson advises when to consider using a Software-as-a-Service model.

Leave a Reply

(Note: This name will be displayed publicly)