Blog Review: March 31

Improving NN on mobile; Python coroutines; PCIe history; medical device security.


Arm’s Pavel Rudko considers several common approaches used to get better performance for neural network inference on mobile devices, such as optimizing and pruning the model and using different processing units to execute different workloads in parallel.

Siemens EDA’s Ray Salemi introduces basic concepts of using Python for verification and how to get Python to talk to an RTL device-under-test using cocotb and coroutines.

Cadence’s Paul McLellan traces the history of the PCIe specification from the early days in 2004 through to the present adoption of PCIe 5.0 and the applications enabled by the upcoming 6.0 version.

Synopsys’ Chandu Ketkar shares five best practices for medical device security that can help defend against attacks on devices and the networks and systems to which they are connected.

Xilinx’s Salil Raje explores how the explosion in unstructured traffic means data center operators are rethinking their current architectures and exploring new configurations that are inherently more scalable and efficient.

SEMI’s Christian G. Dieseldorff sees three consecutive years of record highs in fab equipment spending with a 16% increase in 2020 followed by forecast gains of 15.5% this year and 12% in 2022, with fabs worldwide adding about $10 billion worth of equipment in each of the three years.

Ansys’ Tim Palucka checks out how fluid simulation was used to help develop a water-powered turbine to generate electricity in rural Cameroon, where lots of rivers and waterfalls provide hydropower potential in the rainy season.

Aldec’s Igor Gorokhov demonstrates how to pair thermal sensor data with a visible spectrum camera and computer vision algorithms in a system for automated elevated body temperature screening.

NXP’s Steve Frezon walks through how the company had to shut down its two Austin fabs during the recent winter storm and power outages in Texas, and what went into getting things back up and running again.

Applied Materials’ Jaclyn Kossmann explains why the company’s venture arm is investing in new methods of optical design that replace traditional lenses with micron and sub-micron sized features.

Silicon Labs’ Kevin Smith digs into phase noise challenges by looking at the differences between wander and jitter as well as examining Maximum Time Interval Error, a primary wander metric.

For more good reading, check out the blogs featured in the latest Systems & Design newsletter:

Technology Editor Brian Bailey asks if EDA has failed to innovate, or if the semiconductor industry has shown an unwillingness to adopt new tools and methodologies. As always, the answer is somewhere in between.

Synopsys’ Kiran Vittal explains how to find and fix errors efficiently by using coding assistance for SystemVerilog.

Cadence’s Frank Schirrmeister takes a look at how defense contractors and the US Air Force are increasing verification confidence and deploying digital twins.

Siemens EDA’s Greg Curtis and Ahmed Ramadan, and GlobalFoundries’ Ninad Pimparkar and Jung-Suk Goo, call for enabling a standard simulator-agnostic interface for aging modeling, simulation, and analyses.

Aldec’s Farhad Fallahlalehzari examines why FPGAs were used in the Perseverance rover for applications such as radar transceiver, navigation systems, motor controllers, and computer vision applications.

OneSpin’s McKenzie Ross looks at how a puzzle contest challenged engineers to reveal design secrets hiding in plain sight.

Codasip’s Roddy Urquhart presents a brief history of the RISC-V open ISA.

Leave a Reply

(Note: This name will be displayed publicly)