Date: 28/09/2012
Semiconductor and software: Lot more to come
In the areas of semiconductor chip technology and electronics design automation software (both for running and designing chips), a lot more still has to come. Software is clearly is big enabler of innovation in electronics engineering. Mr. Wesley Ryder, Worldwide Technical Director, Mentor Graphics shares some of the technology and market trends in electronics design. Below is the Q&A.
Q: C is becoming a universal language, in 2/3 years is C completely replacing assembly and HDL?
Wesley Ryder: No, I don’t think so, I spoken this morning, certainly not replacing HDL or Verilog, Verilog is now becoming rather System Verilog, a much more accepted standard here. We did a verification survey a couple of years ago — that shows actually India was leading in terms of percentage-respondents in the adoption of System-Verilog. So, that world is not necessarily in infancy, but still very much long way to go there still, the two communities will co-exist for long time to come, and the tools will co-exist to enable the two for a foreseeable future.
Q: Apart from the programming languages, what are the software engineering changes happening in both VLSI and embedded Systems (the mega changes)?
Wesley Ryder: All about muti-core stuff, that’s where you see a lot of stuff now. From the EDA prospective, the interesting change that we see is the way we are facilitating the two software and hardware work to co-exist more now. We spoke a lot this morning (at the Mentor Graphics’ user conference held in Bangalore) in the key note about the co-design and co-verification, and in the future— we are looking at solutions that allow people to rely much more on their favorite IDE, when they compile, debug and basically drill into the hardware.
Role of open source software:
Wesley Ryder: Yeah! It’s an important thing and is something that Mentor is much involved-in in the software world, and definitely play an increasing role.
Q: In the EDA industry, Synopsys, Mentor Graphics, and Cadence having huge share, do you still see some small players having a role to play in coming years? Or is it going to consolidate further!
Wesley Ryder: I am not sure! Well, the consolidation is inevitable, I think in EDA and as in semiconductor and other industry. To what extent that will reshape the landscape now is not clear, because— if you look at the way things are gone now, you have Mentor Cadence and Synopsys — really the big players, and now with Magma having disappeared, there is very large gap now. Its an ongoing thing, there will always be acquisition of smart new technologies, but those kind of things won’t reshape the whole landscape of the industry in future.
Q. Do you see current trend of using third party IPs to grow further for next few years?
Wesley Ryder: Absolutely, if we look at modern chip-size with billions of ICs, its understood-fact that no one designs now from the bottom-up now. Most designs now are based on derivatives of previous generation or chip involving lot of new IPs. It’s going to be the way to go. IP will also take different form and shape and become more and more complex
On the issue of reliability!
Wesley Ryder: You just have to look at the way markets are growing, I think people accepted, its inevitable Though the issue is depending on the source of the IP, but lot of the vendors are now are very reliable and IP is well tested and you have some emerging standards.
Tested vs not tested IPs!
Wesley Ryder: Its obviously a mixed bag, depending on how you look at it — lot of the IP now very well verified and comes with verification suite along with it, and we have emerging standards around tests to allow IPs to be brought together and tested effectively— to make sure these sub systems work.
Q: In Asia, particularly, India and China going full speed in chip design? Do you see they are (VLSI engineers) on-par with Europe and U.S. in complete product design?
Wesley Ryder: Yeah. It is interesting! I thought about that this week, because a few people were asking me, how I saw India evolving in design now, and I think we all need to— including myself, stop thinking — kind of — India some kind of secondary design region. If we look at the companies that are here, its not a fact that they may are just designing blocks for some parent company based somewhere else. The companies (in India) are basically have 100s of place-and-route chip designers designing smart products. It is no longer the case; the India is kind of any less-than-par than any other region. There are great designers here, some of the products are pretty impressive.
Q: If you look at the present level of processing, its more of a data(text) crunching, if we look at the future its more of complex audio/video kind of a data, huge media streaming of HD type, in that sense how you see future chips going to be?
Wesley Ryder: Six years of time, obviously— we are looking at much finer design rules in the processes, as you said — video/audio kind of applications are very complex and very large design— and the challenge will change a lot. Especially on the video, per example, to execute huge datastream— to ensure all things operate correctly, that’s going to need increasing verification technologies like hardware emulation — those kind of things. And again India is not behind the curve in that. This week I met number of customers doing that, they are already accessing hardware-emulation remotely for video applications, and certainly lot more of that to come up.
Below is the video of the interaction: