Multiple asynchronous clocks are a fact of life on today’s SoC. Individual blocks have to run at different speeds so they can handle different functional and power payloads efficiently, and the ability to split clock domains across the SoC has become a key part of timing-closure processes, isolating clock domains to subsections of the device within which traditional skew-control can still be used.
As a result, clock domain crossing (CDC) verification is required to ensure logic signals can pass between regions controlled by different clocks without being missed or causing metastability. Traditionally, CDC verification has been carried out on RTL descriptions on the basis that appropriate directives inserted in the RTL will ensure reliable data synchronizers are inserted into the netlist by synthesis. But a number of factors are coming together that demand a re-evaluation of this assumption.
A combination of process technology trends and increased intervention by synthesis tools in logic generation, both intended to improve power efficiency, is leading to a situation in which a design that is considered CDC-clean at RTL can fail in operation. Implementation tools can fail to take CDC into account and unwittingly increase the chances of metastability. (more…)
Just before the design automation conference in June, I interviewed Sarath Kirihennedige and asked him about the drivers for clock-domain crossing (CDC) verification of highly integrated SoC designs, and the requirements for handling the “big data” that this analysis produces. He discusses these trends and how the 2015 release of Meridian CDC from Real Intent meets this challenge.
He does this in under 5 minutes! You can see it right here…
On April 19, 1965, Electronics magazine published an article that would change the world. It was authored by a Fairchild Semiconductor’s R&D director, who made the observation that transistors would decrease in cost and increase in performance at an exponential rate. The article predicted the personal computer and mobile communications. The author’s name was Gordon Moore and the seminal observation was later dubbed “Moore’s Law.” Three years later he would co-found Intel. The law defines the trajectory of the semiconductor industry, with profound consequences that have touched every aspect of our lives.
The period is sometimes quoted as 18 months because of Intel executive David House, who in 1975 predicted that chip performance would double every 18 months; being a combination of the effect of more transistors and their faster switching time.
What if Gordon Moore got his math wrong and that instead of the number of components on an integrated circuit doubling every couple of years, he said every three years? (more…)
“Imagine stepping into a car that recognizes your facial features and begins playing your favorite music. A pair of gloves that knows the history of your vehicle from the time of its inception as a lone chassis on the factory floor. “ –Doug Davis on IoT@Intel
Trends in the Internet of Things (IoT) has been fascinating to follow.
In my last blog on the topic I mentioned the 4 challenges facing an IoT system as spelled out by James Stansberry, SVP and GM, IoT Products, Silicon Labs: functionality, energy, connectivity and integration.
Four elements make up successful IoT hardware
This had me thinking… Does this paradigm apply only to the hardware of IoT?
A long-time EDA industry analyst, Gary Smith, passed away on Friday, July 3, 2015 after a short bout of pneumonia in Flagstaff, Arizona. He died peacefully surrounded by his family.
Gary Smith in 1963
Gary was from Stockton, CA and graduated from the United States Naval Academy in 1963 with a bachelor of science degree in engineering. His class yearbook says: “He managed to maintain an average grade point despite the frantic efforts of the Foreign Language Department. Tuesday nights found Gary carrying his string bass to NA-10 practice.” Gary continued to be a musician and played his electric bass for years with the Full Disclosure Blues band at the Design Automation Conference Denali party with other industry figures. The band started out of a jam session in 2000 with Grant Pierce who asked Gary to help put together a group for the following DAC. Gary had suggested Aart de Geus as lead guitar who ended up giving the band its name.
Gary got into the world of semiconductors in 1969. He had roles at the following companies:
LSI Logic, Design Methodologist (and RTL evangelist), 2 years
Plessey Semiconductor, ASIC Business Unit Manager, 3 years
Signetics, various positions, 7 years
In 1994 he retired from the semiconductor industry and joined Dataquest, a Gartner Company to become an Electronic Design Automation (EDA) analyst. Gary described his retirement this way: “instead of having to worry about Tape Outs and Product Launches, I get to fly around the world and shoot off my big mouth (which I seem to be good at) generally playing the World’s Expert role. Obviously there isn’t much competition. Now if I could only get my ‘retirement’ under sixty hours a week I’d be happy.” (more…)
I was an organizer for the industry DAC panel on “Scalable Verification: Evolution or Revolution?” held during the second week of June in San Francisco. While the industry generally agrees on methodologies used to verify IP blocks or subsystems, we lack consensus on approaches required to verify SoC integration and system-level functionality of embedded systems. One of questions addressed by the panelists was “Can existing standards and methodologies be extended to address system-level challenges, or are new approaches required?”
The panel was moderated by that veteran verification technologist Brian Bailey. He was excellent in steering the panelists through this deep topic. The panelists were all from semiconductor companies (not EDA) and included the following:
Ali Habibi, design verification methodology lead at Qualcomm
Steven Jorgensen, architect for the Hewlett-Packard Networking Provision ASIC Group;
Bill Greene, CPU verification manager for ARM
Mark Glasser, verification architect at Nvidia.
Brian has written an excellent article on the panelists insights in his column on SemiEngineering.com. Here are a few quick snippets to entice you in reading the entire piece:
“Simulators are not making effective use of the advances in the underlying hardware. Design sizes are growing faster than the improvements they are making.”
“Design reuse has not helped us, and even if you change only 20% of a design you still have to completely re-verify it. We need to be able to describe features and functionality in an abstract manner, and from that derive the inputs to the verification tools.”
“You might think that we are able to re-use much of our verification collateral from the IP, unit and top levels into the system-level environment, but this isn’t the case. You can’t find new bugs by running stimulus that was used in the past, and this means that the notions of coverage are different.”
You can read the entire column at this link: Wrong Verification Revolution Offered. Be sure to add your comments at the end and let Brian know what you think is missing in verification.
Richard Goering at his 30th DAC, San Francisco in 2014
Richard Goering, the EDA industry’s distinguished reporter and most recently Cadence blogger is finally closing his notebook and retiring from the world of EDA writing after 30 years. I can’t think of anyone that is more universally regarded and respected in our industry, even though all he did was report and analyze industry news and developments.
When Richard left EETimes in 2007, there was universal hand-wringing and distress that we had lost a key part of our industry. John Cooley did a Wiretap post on his DeepChip web-site with contributions from 20 different executives, analysts and other media heavyweights. Here are a just few quotes that I picked out for this post: (more…)