Tuning into Jim

I recently attended the annual TSMC Technology Symposium in San Jose, at which TSMC’s Vice President of Design and Technology Platforms, Dr. Fu-Chieh Hsu, presented a talk on, “The Green Design Concept.” Dr. Hsu proceeded to describe the world’s current power needs – a staggering 300 trillion Watt-hours/day – and how the silicon-based devices in the computing, communications and consumer markets, while only representing a small fraction of that number, still account for over 1 trillion Watt-hours/day, a number that is rapidly growing.

Dr. Hsu continued by explaining how both the semiconductor manufacturing and design communities can help minimize the power needs of silicon-based equipment. On the silicon foundry side, technology and production quality along with environmentally friendly production techniques all contribute to minimizing energy use during chip production. For their part, chip designer can develop energy-efficient designs in areas such as fuel-efficient vehicles, smart office management systems, and intelligent lighting and appliances in the home. There is nothing really new with these ideas. However, designers can also do a lot to minimize “silicon waste” through techniques such as minimizing on-chip logic, design margins, timing slack, and circuit area, and maximizing the use of silicon-proven IP. Naturally, these techniques require good design tools and accurate models from and close cooperation with EDA vendors, along with partnerships with semiconductor foundries, to develop “just right” designs instead of “over designs.” It also means that chip designers have to accept a new way of designing chips, minimizing the over-design techniques that they feel are required to provide the safety margins necessary for acceptable yields. The benefit to the designer is a smaller, more efficiently designed chip and higher profit margin.

The concept of “Green Silicon” is a good one and a concept that we can all embrace. Yes, it will take significant effort on everyone’s part, but isn’t the end result – better silicon power efficiency and more profit on the bottom line – worth it? As Dr. Hsu stated at the closing of his presentation, “Green Designs lead to Greenbacks.”

My wife got me a very nice Christmas gift – a high-end Plantronics Bluetooth headset for my cell phone.  When she told me where she bought it (Radio Shack), I checked the price, which was $100.  I then went on the Web and “shopped around,” finding the same headset for as low at $60.

My next task was reviewing the headset’s features on-line (I still had not opened the box).  Lo and behold – it had a lot of nice capabilities, including many that I would never need or use (think Microsoft Office and how many of the capabilities in Word, Excel and PowerPoint you really make use of).  Over the next couple of days I looked around for a headset with the features I did need, at a more reasonable price, finally buying an $80 Jabra unit that, with rebates, was only $30 at Fry’s.  Finally, I returned the Plantronics headset to Radio Shack.

The purpose of this story is not to show what a great shopper I am (that point is debatable), but to impress upon you the importance of buying only what you need in a product, not what someone else (spouse, friend, salesman) thinks you need.  Unless, of course, you like to waste your money.  This philosophy applies to virtually any purchased item, but is particularly relevant to electronics products and their rapidly growing feature sets.  Whether in the market for a new computer, flat screen TV, cell phone or music player – determine what you really “need” and then add on your most important “wants.”  This should guide your purchase decision, not the bells and whistles you’ll never ring or blow.

Oh, by the way, when I discussed my headset ideas with my wife, she agreed with the decision to buy the far less expensive unit.  More money for us to spend on other things was probably in the back of her mind.  Smart woman!

The other day I was watching a segment on 60 Minutes about the Geek Squad, the company formed in the mid-1990s to help consumers set up computer, home-theatre and other electronic systems.  The premise was simple – most electronic systems in the consumer space are too complicated for the average person to tackle successfully.  Remember the joke about all those VCRs blinking 12:00 because no one could set their clocks?  If you think things were difficult then, look at what we have now.

The common themes in consumer electronics are – add more features, shrink the footprint, decrease power consumption and reduce prices over time.  Vendors have done these jobs very well.  However, along the way they forgot to do one thing – make it easy for the average consumer to buy an electronic device and be able to set it up and use it WITHOUT OUTSIDE HELP.  I’ve been immersed in Silicon Valley electronics for over 30 years and still have to really put my mind to it when I buy a new gizmo.  Imagine the plight of a middle-aged gas station attendant somewhere in a small town in Iowa.

Both electronics vendors and their customers – you and I – need to place more emphasis on the human interface – making it easier for the average person to set up and learn to use an electronic device, be it an HDTV, MP3 player, digital camera, or cell phone (to name just a few).   Low prices are nice, but how many of us really use all the whiz-bang features in your new electronic “toy” (or even in your version of Microsoft Office)?  Make ease of learning and ease of use two factors to consider when shopping for a new phone or camera.  Consider what features you really need (not what sound impressive) and then check out the competing products with that feature set.

Successful vendors of consumer electronics should be (and sometimes are) the companies that really do consider the ability of a potential customer to deal with their products.  Consider the iPhone.  Its success, I believe, has been based as much on its easy-to-use and understandable touchscreen and virtual buttons than on what you can do with it. 

Product understanding and ease of use are the real keys to consumer electronics differentiation.  If more of us kept this in mind while shopping, then maybe more of the vendors of these products would wise up.

In a recent opinion piece in the electronics trade press, a well known editor printed a very controversial picture of the semiconductor IP industry, going so far as to say, “It’s simply a lost cause, with a questionable business model.”  The article created an outpouring of responses from IP vendors and many other members of the semiconductor industry, some going directly to the editor but many amongst the IP community members.

Whether or not you agree with the opinion piece, the article has caused people to consider about what is right and wrong with IP and to discuss what can be done to improve a situation which most people think is far from perfect.  The IP segment of the chip industry is unique in that there are hundreds of IP vendors offering everything from simple library cells to very complex video processing and encryption cores, the latter which would have been sold as separate chips just a few years ago.  The problem with such a complex ecosystem is that there is precious little commonality in the way IP vendors configure and sell their products and IP integrators evaluate them.  A first step in dealing with the qualification issue was taken by the VSIA with their Quality IP metric (QIP), but their work on this and other IP technical issues is being passed on to other standards organizations, including the IEEE, with the dissolution of the Alliance.  The business issues of pricing (getting value into the work put into developing the IP) and selling continue to bug the industry.

Getting people to talk about IP problems is a good thing and some solid results have already come about.  In a few weeks, the publisher of the aforementioned IP article will be hosting a panel to discuss how serious IP problems really are and some ways of overcoming these problems.  Susan Cain and I have also been approached by many companies and IP consultants about forming an IP industry organization that brings the IP companies together to discuss and find solutions to business and trading issues – EDAC with EDA tools and the FSA with foundries and fabless semiconductor companies have shown the value of such unions.  Such an organization makes a lot of sense, since the need for semiconductor IP will continue to increase as process nodes and time-to-market cycles continue to shrink. 

It behooves the chip industry to tackle the IP segment’s problems now in order to enable the continuing growth of the semiconductor market.  Talking about these problems, even if the talk was instigated by a controversial article, is a good first step in this direction and, hopefully, will lead to additional activities to ease the pain of IP integration.

The next time you leave your home, look about you as you travel.  What do you see?  Cameras – lots of cameras –  on traffic lights, at toll booths, at the ATM, in your local gas station and just about everywhere else.  Like it or not, someone is capturing your image several times a day for a variety of reasons, usually associated with security or safety.

Welcome to the Age of Video Surveillance.  The world in which we live, play and work has become a lot less safe than it was just a few years ago and we, willingly or otherwise, are giving up some of our privacy to feel a little more comfortable with our surroundings.  So, what does this mean to the semiconductor industry?  An exploding market offering a wide range of new product opportunities.

Market research firm iSuppi has the hard numbers.  They expect global video surveillance camera revenue to grow from $4.9 billion in 2006 to more than $9 billion in 2011; unit shipments of video surveillance equipment to more than double from 29.8 million in 2006 to 65.7 million in 2011; and the market for surveillance-camera chips should hit $1.25 billion in 2011.

If you are a processor vendor who targets video processing applications, video surveillance should be in your crosshairs.  Likewise for the companies who develop video analytics, sometimes called intelligent video, algorithms – software that surveillance systems use to analyze a scene and flag it if it represents a potential safety or security risk.  CMOS sensor vendors, VCR manufacturers and video camera companies also are potential beneficiaries of the huge growth in video surveillance applications and equipment demand.

Video surveillance may not be the next semiconductor “killer app” but it does represent a very real opportunity for chip vendors and system houses.  And when you’re out and about, make sure your hair is combed, your clothes pressed, and smile – someone is probably looking at you.

I have a confession to make.  I recently returned from a 10-day vacation in the Caribbean and Miami.  You will be amazed to hear that I did not take my laptop and did not access email the whole time – and survived the experience!

Yes, for the first time in two years I traveled away from my office without my laptop computer in tow and the world as we know it did not end.  I can hear you think, “I am shocked, shocked that someone would be so irresponsible.” But, it really felt good not to feel obligated to “check in” several times a day to see what was in my In-Box.

One cannot do such a thing blindly, of course.  First – make sure that your colleagues are aware of your obligations to clients, customers and so on during your absence and provide them with your contact information “just in case.”  Second – plan for your absence by wrapping up as many loose ends as possible before you leave.  Finally – be prepared for a flood of email messages upon your return.  I had over 3560 “good” emails and more than 600 in my junk folder.

The point of this monolog is to show that there is life without 24/7 connectivity (I had my cell phone off most of the time as well).  We Silicon Valley types are trapped into thinking that we must always be available to others anytime, day or night.  This is bad karma and, for the most part, isn’t necessary.

I had a wonderful, relaxed time being isolated from work and recommend the experience to everyone.  A vacation should be just that - leisure time away from work devoted to rest or pleasure – and, darn it, I think we all need to experience it that way.   

Let me know whether or not you agree with me.

Having had a week to catch up with my normal work after returning from DAC, it’s time to discuss a few impressions of the conference and show.

First – what has happened to the DAC attendees?  It was pretty obvious to me that there were a lot fewer people at this year’s conference than at DAC 2006.  Last year, with DAC in late July, a scarcity of European attendees was understandable (holiday time on the Continent).  However, the decrease in overall attendance this year included, again, fewer Europeans than traditionally show up.  The European press was, again, mostly absent this time around – maybe having DATE in Nice six weeks prior to DAC was not a good decision (although I understand the weather was beautiful).  I found show floor attendance sparse most of the time and other people mentioned to me that some of the technical sessions they attended also had a smaller audience than they expected.

The choice of venue for DAC – the San Diego Convention Center – is a very good one.  Many hotels are in reasonable walking distance, close enough that shuttle buses were not needed.  The downside of the SDCC is that several meeting rooms upstairs are a relatively long walk from the exhibit floor, meaning it took longer to corral attendees for events in those rooms.

If you think all exhibitors come to DAC to meet new customers, think again.  Several companies with whom I talked told me that they use DAC as a central, pre-arranged meeting place for customers and potential customers.  For several of these companies, their meeting rooms were booked solid prior to DAC’s start, and finding and meeting with new companies was a secondary goal.

Finally, there definitely was something missing from DAC this time around – the “buzz” that usually accompanies an important conference.  People walking around were mostly subdued and even the exhibitors didn’t seem to have the same fire that I am used to seeing at this show.  Maybe it’s time for the DAC powers to be to evaluate how they run the event and make some changes that may re-invigorate what seems to be a tired conference.

Part of the fun of attending DAC every year is going to several of the events that are not “DAC sponsored,” in other words, events that are put together and sponsored by companies, analyst groups, standards organizations and the like.  Very often these turn out to be some of the highlights of the “DAC experience.”  The following list contains a few of the unofficial DAC activities that caught my eye and I plan on attending.

·         EDAC Executive Reception, Sunday 5-7PM at the Marriott.  This was originally the Sunday Dataquest briefing, the unofficial opening of DAC, we all loved for networking and informational purposes.  Register at http://www.edac.org/invites/dac_reception_07.htm?opendocument

·         Gary Smith (now at Gary Smith EDA) is back the same evening, also at the Marriott, from 7-8:30PM, to discuss EDA trends and numbers.  Register for this at https://app.e2ma.net/app/view:Join/signupId:20318

·         VSIA, the IP standards alliance, has a luncheon on Tuesday from noon to 1:30PM in Room 32AB on “IP Quality: How to rate IP using VSIA’s QIP Metric standard.”  Mentor, Denali and LSI will all make presentations and you can register at http://www.vsia.org/events/DAC2007Registration.htm

·         Tuesday will feature the always popular Denali party – nothing else need to be said about this except to register for your (limited supply) ticket at https://www.denali.com/docs/servlet/denalisoft.dac.Rsvp 

·         MOSAID and Sidense are sponsoring a very interesting panel lunch on Wednesday from Noon until 1:30PM in Room 29CD on “Solving the SoC Memory Puzzle.  The moderator will be the knowledgeable and entertaining Ron Wilson, EDN’s executive editor.  Register on the MOSAID website at http://www.mosaid.com/corporate/products-services/ip/dac_register.php

·         Accellera, the standards committee that works on design languages and methodology, is hosting a breakfast and panel discussion on Wednesday morning, 7:30-9AM in Room 26AB on IP: The Next Frontier for SystemVerilog.  Sponsored by Denali, you can register at http://www.accellera.org/events/register

A couple of weeks ago, EE Times reported that market research firm Semico announced that Structured ASICs were still strong and had the potential for “explosive growth” in the future.  This surprises me – a lot!

Despite a promising beginning, and originally backed by several major semiconductor companies, Structured ASICs have never reached the levels predicted by many analysts three years ago.  There are several reasons for the less-than-expected growth in this segment, including non-agreement of what comprises a Structured ASIC; greatly different architectures and technologies offered by the professed Structured ASIC vendors (some FPGA-like, some ASIC-like and some in-between); pressure on the high-end from ASICs with new, lower cost business models and on the low-end from FPGAs; and lack of tool support from the major EDA companies.  Then came the pullout of chip vendor LSI Corp. (then LSI Logic) in early 2006, quickly followed by tool vendor Synplicity, from the Structured ASIC market.

So – why the optimism of Semico in the future of Structured ASICs?  It may (at least partially) be due to, as EE Times states, Semico creating a definition that it said ‘captures most of the important features’ that can be used to define a structured ASIC.  The EE Times article then goes on to list what these features are and, upon reviewing them, they sound to me a lot like what comprises an ASIC or ASSP.  With this definition, Semico predicts a compound annual growth rate of 28.2 percent in the years 2007 through 2011. How “real” is this growth?  Well, by Semico’s current Structured ASIC ‘definition’ a lot better than it would be based on what I remember as defining a Structured ASIC.  Me – I’m not nearly as optimistic about Structured ASIC’s future.  The market segment may not be dead, but it currently sure doesn’t have a high quality of life. 

That’s my opinion – let me know what yours is.