Graham is VP of Marketing at Real Intent. He has over 20 years experience in the design automation industry. He has founded startups, brought Nassda to an IPO and previously was Sales and Marketing Director at Internet Business Systems, a web portal company. Graham has a Bachelor of Computer … More »
CEO Viewpoint: Prakash Narain on Moving from RTL to SoC Sign-off
September 19th, 2013 by Graham Bell
In early summer, Real Intent was among companies at the forefront of promoting a more unified approach to RTL sign-off, unveiling a deal that linked its market-leading Meridian CDC clock domain crossing suite to design-for-test tools from DeFacTo Technologies. Now though, as president and CEO Prakash Narain explains, the company is focusing on the broader concept of SoC sign-off.
“Take CDC as an example,” explains Narain. “You cannot sign off on CDC purely at the RTL because it is very possible that synthesis will introduce problems that cause your design to fail. So, you need the capability to run a small part of the CDC verification on your netlist as well.
“We provide that with Meridian and that is why we see ourselves as an SoC sign-off company. You are often looking at things that cannot be done at just the RTL level.”
It is a model that fits RI’s emphasis on static verification very well. The company’s tools address particular failure modes, including also areas such as X propagation and timing constraints as well as the established Ascent Lint family.
The SoC sign off strategy means that your tools must either catch issues early or monitor for them wherever they may occur. And as they do that, in as predictable and thorough a way as possible, Narain says that you help address the main challenge facing front-end design.
“What is driving front-end verification largely comes down to making methodologies as efficient as possible,” he says. “They are becoming more and more complex because the designs themselves are becoming larger and larger, but they need to be done in the same time as the ones before.”
Tools aimed at particular design challenges deliver this efficiency by allowing some aspects of verification to be accomplished in parallel.
“You can think of a methodology where everything is done in simulation, but we all know that is going to take a long time, too long,” says Narain. “Static verification allows you to develop a methodology where verification for a specific failure mode can have a complete process associated with it. And that process can be run alongside those parts of your flow that still have to use other techniques.”
A political analogy is useful here, he adds, “You can borrow [former US Defense Secretary] Donald Rumsfeld’s terminology about ‘known unknowns’ and ‘unknown unknowns’. For known unknowns, static solutions are the best option. For unknown unknowns you can use techniques that offer much broader coverage, such as simulation.”
There has been some discussion around this distinction about the possible impact of a time-to-market driven shift to probabilistic computing techniques, but Narain is doubtful that this will diminish the power of static verification.
“You need to look at how design teams are managing things, at why static solutions are getting more and more interest,” he says. “Managers are not adopting risky models. With probabilistic models, they are going to be very worried that as designs continuously increase in size, so does the probability of some failure occurring somewhere. And it just needs one failure to force a respin.”
Coming back to the integration of static verification within the broader SoC sign-off concept, the vendor must address other needs. As the DeFacTo deal showed and, more recently, Real Intent’s entry into Synopsys’ InSync program, the tools need to play nice – interoperability or bust.
“Customers want maximum predictability and maximum control over their methodologies, so the more integrated you are with other front-end workflows, the greater the efficiency you will offer,” says Narain. “That makes these partnerships very important.”
But the other important factor is that the static tools themselves are constantly refined. In that regard, the recent release of Meridian 5.0, with a hierarchical flow headlining a number of efficiency-driven innovations, is a good example. And, says Narain, there’s more to come.
“It’s easy to create something that doesn’t add value over the existing flow. We need to continuously increase the ROI users derive from these solutions,” he argues. “Our job is to make people think less about having a particular technology gap to fill and more about how static solutions improve and enhance flows.”