May 05, 2003
Design to Silicon
Please note that contributed articles, blog entries, and comments posted on EDACafe.com are the views and opinion of the author and do not necessarily represent the views and opinions of the management and staff of Internet Business Systems and its subsidiary web-sites.
Anon in Marketing - "Just got through reading this [article]. So this is about the trip you took with your son, right?"
(Editor's Note: Our daughter was looking at the PhD program in Computer Science (AI) at Michigan. She has since decided to attend MIT and will enter in the fall.)
April 28th - Closing comments and kudos from the University of Toronto
Kim Hailey - "In your [newsletter] today, there was one very little mistake. HSpice is currently a Synopsys product. Meta-Software merged with Avanti, which merged with Synopsys. Thanks for the great newsletter."
(Editor's Note: I would suspect that many people continue to think in terms of "Avanti tools," even as all of us know that Avanti is now part of Synopsys.)
April 28th - Industry News - Tools & IP
Tony Larson, Vice President of Marketing for Verplex Systems - "Thank you for your mention of Verplex's new Conformal DP in this week's EDA Weekly. I'd like to clarify a small, but important point about Verplex's new datapath verification product: Conformal DP is designed to verify the functionality of complex optimized datapath circuitry produced by datapath and synthesis tools. We don't verify timing, which is what the paragraph currently reports."
(Editor's Note: This is an important distinction. My apologies to Verplex. The April 28th issue of EDA Weekly has been corrected and is now available on-line.)
Anon, ESNUG Reader - "[Here] is the challenge with NDA agreements. The same term and approach have come to cover a variety of business activities to the point that, if a customer or prospect isn't briefed 'under mutual NDA,' they feel cheated. (Why else, all of those high priced enclosed suites at DAC, for example?)
The thing that makes ESNUG such incredibly valuable reading to me is not that I can learn what vendors' plans are - though, if you don't like science fiction you shouldn't read business plans and vendor roadmaps - but that I get a very accurate picture of customer pain (even agony and, in some cases, near death experiences) that allows me to understand where the opportunities are and where the vendors will ultimately go.
Technically John asked 'See anything interesting under NDA? What did you think of it?' I suspect the answers he will get back are Vendor XYZ showed a compelling solution to speed up simulation by a factor of 10, or Vendor ABC showed us property checking that scaled to 6000 bits of state space.
The net impact will be for prospects to ask Vendor XYZ or ABC for their presentations, which is what the vendors want anyway - buzz.about something they have that's compelling, or at least interesting. There's something implicit in [John's] question that any good reporter or marketing person asks all the time at a conference: 'What have you seen that's interesting?'
Since most presentations are under NDA, it's implicit that what you saw may have been under NDA, and very likely anything that was interesting was under NDA. [John] could have left NDA off of this question and gotten the same answers. I believe that you should honor every NDA that you sign, but most expire pretty rapidly as the information comes into the 'public domain' of a small user community."
April 28th - Cooley twice over - Letter No. 2
John Sanguinetti, CTO at Forte Design Systems - "The answer to Cooley question No. 2 is Magma."
Tim Reeves, Matsushita Electronic Materials, Inc. - "The answer's right there when you click on the red page. It's some company called Magma."
Daniel Payne, Product Marketing Manager at Mentor Graphics - "This sounds like Magma, because there's a press release today saying: 'The end of synthesis as we know it is here,' said Venktesh Shukla, senior vice president of Marketing and Business Development for Magma Design Automation."
Anne O'Neill, Executive Director for the IEEE Solid-State Circuits Society "Letter No. 2 wasn't that hard. The URL of the sponsoring firm was in the image, if not linked. They are www.magma-da.com."
Gloria Nichols, Principal at Launch'M - " FYI, John posted later on his site that his readers told him that Magma was behind the ad."
Sean Murphy, MegaPath Networks - "If you go to the website, it's clearly Magma. It must have been updated since the question was posed. Take another look. That's the problem with living in Internet Time. The half-life of information accuracy is measured in hours."
John Cooley, ESNUG - "To answer your question in Letter No. 2, it was Magma just doing some pre-DAC marketing hype. When [Richard] Goering asked me about the Magma synthesis announcement, I laughed because it wasn't news. Magma had announced that they had synthesis when they first came to market 2 years ago. Nobody's ever written to me about using it in all this time, so I guess Magma's just trying to rustle up some interest in it this time around at DAC. I guess y'all got to have something to talk up at DAC, or the customers will walk on by. :)"
(Editor's Note: Perhaps the clever folks at Magma should write a White Paper entitled "Stealth and Cunning in Creating Buzz," or "Getting Editors to do Your Marketing Work for You.")
Week of April 28th - A query on the current status of SystemC
Cliff Cummings, Sunburst Design - "I think Gary Smith is right about Synopsys' Behavioral Compiler. ("Forte's Foray into EDA's Failure" by Alex Romanelli Electronic News, 4/29/2003) A great tool for certain applications, but way over sold. I was wondering if you have any opinion about SystemC suffering the same problem?
SystemC was oversold as the architecture-to-gates solution with free GNU C simulators. It didn't take long to realize that SystemC was terrible at gate-level simulation, and not too much longer to figure out that it does not compete well with Verilog and VHDL at RTL design and simulation. What does that leave? Architectural analysis and high-level testbench development. In the mean time, engineers, like myself, found good reasons to disbelieve anything positive that was reported about SystemC. I assure you, if we all saw three magnitudes of simulation-speed improvement, we all would have looked much harder at this new tool.
The next question is, since SystemVerilog is going to have a seamless C-interface, do we need SystemC at all? Will SystemC be the preferred interface to SystemVerilog or will designers just use plain old C? It is too early to answer that question, but I am seriously questioning what SystemC adds that I can't achieve with plain old C. Only time will tell.
The SystemC crowd touts the architectural capabilities of the language, but nobody will share what cool architectural things they are doing with SystemC. This leaves engineers like myself asking, 'Where's the beef?'
Maybe some of your readers could point us to some cool SystemC architectural models to convince us, the doubting types, that this really is a good thing. One lesson we are learning is, don't over sell your tools to a bunch of bright designers. They tend to be an unforgiving lot!"
You can find the full EDACafe event calendar here.
To read more news, click here.
-- Peggy Aycinena, EDACafe.com Contributing Editor.
Be the first to review this article