Blog Review: June 1

Photonic PDKs; non-maskable interrupts; data center disaggregation.


Analog Photonics’ Erman Timurdogan, Ren-Jye Shiue, and Mohammad H. Teimourpour, and Ansys’ Bozidar Novakovic, Ahsan Alam, and Peter Hallschmid consider the development of photonic process design kits and the importance of choosing a laser model that can optimally satisfy often conflicting requirements between the number of known laser parameters, the model accuracy, and the computational time.

Arm’s Christoffer Dall explains non-maskable interrupts, including how operating systems software use them and why they are called non-maskable when there are several ways to mask them, as support for them is added to the Arm A-profile architecture.

Synopsys’ Manmeet Walia and Manuel Mota look at how bandwidth-hungry applications are changing data center architectures and find that data center disaggregation, where resources are separated in different boxes and connected optically, is a path toward even more efficient processing of increasingly massive workloads.

Cadence’s Paul McLellan listens in on a panel to find out how systems companies are driving chip design and the challenges involved in building bespoke silicon compared to market silicon.

Siemens’ Katie Tormala shares ten tips to maximize confidence in component temperature prediction to aid thermal design for reliability and efficiency.

SEMI’s Serena Brischetto chats with Guido D’hert of Accenture about the challenges semiconductor companies face in meeting sustainability goals and how to apply analytics, new green technologies, and circular design.

Plus, check out the blogs from the latest Systems & Design newsletter:

Technology Editor Brian Bailey contends that significant improvements in verification are possible, but getting critical mass behind them will be difficult.

Cadence’s Frank Schirrmeister looks at how advances in computing and connectivity are transforming industries from consumer electronics to aerospace and defense.

Movellus’ Aakash Jani warns that for AI compute, simply throwing more chips or processors at the problem is not a scalable solution.

Siemens’ Matt Walsh investigates bridging the chasms separating product engineering teams from component supply and electronics manufacturing.

Synopsys’ Ajit Sequeira finds that parallelism is the only scalable solution for making any application faster.

Codasip’s Rupert Baines explains why licensing models for IP using an open ISA can get confusing.

Renesas’ Michael Joehren suggests a way to get TWS earbuds powered up quickly and in a small form factor.

Leave a Reply

(Note: This name will be displayed publicly)