January 28, 2008
Fireside Chat: Rick Lucier & Jim McCanny
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.
Question – Do you think this situation is one that could have been anticipated 10 years ago?
Rick Lucier – Even 10 years ago, people were pushing for hardware/software co-design, but now I think the tools are ready for it. We have a ways to go still, but the market is definitely there. It has come to the realization that there are a lot more software people out there than hardware people, plus there’s a lot more software in electronic products than there was 10 or 15 years ago. So yes, people could see it moving in this direction 10 years ago, but software development wasn’t the critical bottleneck that it is today – returning, again, to that idea of the bottleneck game.
Look at the iPhone. Apple re-deployed many software engineers from other projects to ensure a timely release of the iPhone. That was all about the software issue, not the hardware. And that’s true today for lots of wireless and consumer vendors. Certainly that’s what our customers are seeing, and they’re building methodologies to address the hardware-software challenge. For big companies, what was an R&D project a few years ago is now a formal corporate initiative to develop methodologies to address this area.
Question – Rank the targets for design in order of importance and/or difficulty at 65 nanometers: Area, Timing, Power, Leakage, Signal Integrity
Jim McCanny – That can’t be done on an absolute scale – the ranking of these targets always depends on the market for the end product. [Having said that], signal integrity is usually a lower priority than timing.
your question, ranking the design targets depends on the vertical.
Question – Are the hardware guys providing a lot of CPU power and the software guys needing to catch up – in particular, a reference to multi-core products coming off the line today.
Rick Lucier – I think they are, but it’s challenging. Clearly performance on processors isn’t growing like it used to, although definitely there’s more horsepower via multi-cores.
Utilizing multi-cores depends on the application. Some application fit a multi-core, but it requires a fair amount of retooling. The question is, where do we take advantage of multi-cores in the software stack? Do you solve it at the application level? At the middleware level? Seems like a real opportunity exists to solve it in the middleware and avoid the retooling at the application level.
There’s an opportunity there to do all of this, but it’s hard to look at every application [and develop a solution]. You have to ask, can I parallelize it at all? The answer may be gated by the design data, which may not lend itself to parallelization. So, yes – maybe you could say the hardware guys are ahead of the game, because the cost of “porting” remains quite high, without a general scalable solution.
Jim McCanny – No matter what people may say, the hardware guys can’t just say to the software guys, Here’s your hardware. Now where’s the software?’ It’s a lot more complicated than that. It requires the software guys figuring out how to make the software run faster on the new hardware. [Overall], I think the hardware guys haven’t had enough conversations with the software guys: How are you going to use this? What are your end applications and how will they benefit from this?
Question – How do we re-position the EDA tools to run on multi-core processors?
Jim McCanny –The problems in EDA are not simple, although the development systems are now very cheap. As recently as 5 years ago, you needed a whole campus of servers to [work through the EDA algorithms], plus a whole bunch of PhDs. Now the hardware guys have given the software guys so much CPU power, that there is a lot of potential for new solutions [on the horizon], although it will take a decent amount of time to implement it all.
can put a lot of work into it and not realize a payback.
Question – Where is Ground Zero today for EDA tools development? Has it moved from North America to places like India and/or even China?
Jim McCanny – It doesn’t really matter where ’Ground Zero’ is, because I still believe the [bulk of the innovation] comes from startups wherever they’re found. The [folks in startups] are not held up by the baggage of having to support the existing systems [that characterize the Big EDA vendors], or the baggage of the management of the larger organization pulling them in multiple directions.
That’s why the little startups in EDA are so vital to the survival of the industry. And, that’s not to say that the Big Guys don’t come up with ideas, or that every startup is innovative.
Rick Lucier – By and large, development is still centered here in North America, but a fair amount of development has moved offshore – particularly to India, China, Russia and Poland. It’s really all over the place these days as people [continue to] take advantage of where the talent it. The situation is definitely more fluid.
Not so many years ago, people were rushing to do things offshore, and today people continue to push things offshore, but they’re also bringing things back depending on the stage of development. I’d say my gut answer to the question is, Ground Zero is still in North America.
Question – If it’s best in EDA to be near the customer and the customer is moving offshore, doesn’t that push for things to move elsewhere?
Rick Lucier – Well, the customer is everywhere today. It’s even more widespread than just saying the customer is globalized’. So, you move offshore if you have to be there. In the past, it’s been purely a question of cost and talent, but now it’s cost, talent and the need to be closer to the customer. It just makes sense to have a fair amount of representation. Certainly in India, where we have a fair number of multi-nationals, it’s not just a cost thing today. It’s a question of wanting to be closer to the customer.
Question – Should people be concerned about consolidation in the EDA industry? Is the customer served by one-stop shopping?
Jim McCanny – If what you’re asking is: Are people standardizing on one supplier? The answer is yes. That’s the case, at least, on the surface – but I think it’s more complicated than that. If you need to get a design done, you’re going to go to a specific [tool vendor] to get a specific answer. But, in some major areas that are traditional, particularly in the digital design flow, you’re starting to see there’s not a whole lot of choice between the players, so the customers are just looking for discounts. That’s good news on the buyer’s side, although the solutions may not [be all one could wish for].
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