Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
Zona Zócalo: All that Zazz
August 28th, 2009 by Peggy Aycinena
The Zona Zócalo in Mexico City is the center of action in town. EDA startup Zocalo Tech wants to channel that same energy and become the center of action in Assertion Based Verification. Hence, they’ve recently announced their flagship product, Zazz.
Per the company, “Zazz makes using assertion libraries quick, easy and accurate by automating tedious error prone task with its easy to use, intuitive GUI. With Zazz, the checkers from the most widely used assertion libraries can be attached to a design and documented in minutes.”
In conjunction with the Zazz announcement, I had a chance to chat by phone with Zocalo President Howard Martin. For starters, I asked Martin to give me an update on Assertion Based Verification.
Martin said, “ABV is considered to be the future of verification, but the acceptance has been pretty slow nonetheless. Recently, I looked on John Cooley’s site and found his report from the 2004 SNUG meeting where [Synopsys CEO] Aart de Geus was talking about assertion libraries. De Geus maintained that assertion libraries were going to move the industry forward.”
Martin said that optimism still stands, but added, “There’s a bottleneck that continues to keep ABV from widespread acceptance. Assertions are difficult to create and difficult to reuse – the biggest problem in ABV. Assertions have to be documented, which takes a lot of time. And, there are additional problems depending on whether you‘re looking at the process from the point of view of the designer or that of the verification engineer.
“For the designer, we’re talking about assertion libraries, but libraries are hard to use. Even though Cadence, Mentor, and Synopsys all have libraries that cover up to 100 percent of a designer‘s requirements, they’re not being widely used. Accellera’s OVL [Open Verification Library] is supposed to be the big answer for creating assertion checkers, but it’s also not being used a lot.
“Instead, people are pushing for designer-provided assertions, but they’re not in widespread use either because it takes a big dent out of a designer’s time to create them. Designers won’t put assertions on a design if it impacts their design time. It’s just too hard. So neither libraries, nor designer-provided assertions are being widely used. Zazz is our solution. It takes these particularly tedious, error-prone tasks and automates them.
“There are approximately 50 different assertion checkers widely used today. With Zazz, we can do any of them in minutes – create them, attach them, create the line statements [in the code]. In fact, of those 50 checkers, 10 of them represent up to 90 percent of the assertions, and we can do those 10 in just 1 minute. Again, everybody agrees that designers should add design checkers, but if it’s a problem to add them, there will be no assertions checkers to check for.”
Martin also explained the issues that plague verification engineer: “If the designer has decided on ABV, the verification engineer has to do the best he can to relate to the assertion checkers. But if the library only cover about 10 percent of the verification engineer’s requirements – if his assertion checkers are at a higher level, at the interface between the block level and the system level, for instance – again there are problems.
“Zazz helps the verification engineer by taking an assertion library and encapsulating it with a GUI that allows him to take the documentation and [facilitate the process]. The result is a great improvement in productivity, and therefore a lower price of design and verification.
“Plus, our tool is not expensive. Our initial offering has a list price for a one-year license of only $5000.”
Clearly Martin is jazzed about Zazz, which prompted a question about the name. He said, “We chose the name Zazz, because it means something that’s outstanding and unique. Something that, the more you have of it, the better you are!
“Nobody so far in EDA has actually been attacking the problem we’re solving at Zocalo with Zazz. In fact, this technology only became an obvious opportunity to us about a year ago. Now we’re happy to say we have our infrastructure in place, the website, and the complete product offering. We’ve completed reams of testing and demos on demand, and really believe we’re bringing something unique to the marketplace – something that will be at the center of ABV.
“There are over 30,000 designers and verification engineers at work around the world today. With Zazz, we can help move the design process up the ladder to the testbench through automatic documentation, plus lessen the difficulties of using SVA [SystemVerilog Assertions]. Previously, SVA has been really tough to use and, as a result, reusability has been zilch.
“If a designer wants to reuse an assertion checker, he needs to go in and cut-and-paste, but the process is such that it’s way too easy to make errors. As a result, designers prefer to use their old assertions and then start over again with their next design.
“With our Zazz technology, however, we’ve hidden the complexity of this type of reuse. We keep track of modifications in the assertions and only update what’s been changed. We also format everything into a structured verification plan, which from the point of the view of the verification engineer is definitely a productivity booster.
“No matter how you look at it, Zazz is very cool!”
Given his enthusiasm, I asked Martin if he thought Zazz could be the next hot product in EDA.
He was quick to reply: “Absolutely! It’s a great, clean product and doesn’t require any interfaces. It supports both versions of Verilog, supports OVL, plus assertion libraries from all three major EDA companies. Zazz will definitely be hot!”
Okay, there it is. Consider yourself been fairly warned. Zocalo Tech has every intention of becoming the center of action in town!