Design: July 2009 Archives

"Hello, it's your fridge here. You need more milk. And, have you ever considered emptying out that yoghurt? It's getting a bit rank in here."

Yes, it's the return of the internet fridge. Now powered by Android.

Mentor Graphics, which sells tools for embedded systems, has bought small startup Embedded Alley Solutions, which specialises in Linux and Android tools. Mentor normally buys a company during the week of the Design Automation Conference (DAC) and this year is no exception. But if you were expecting Magma Design Automation to be the target. Well, that didn't happen.

The plan is to promote Android as software that can be used in home appliances, information displays in cars and medical systems. In reality, the fridge will probably remain as Internet-free as it has done for the past 50 years. But there are other embedded systems that could make use Android.

By running Mentor's existing Nucleus operating system on one or more cores in a multicore processor, Android can take care of a lot of the user-interface functions while the Nucleus bit looks after the real-time stuff. Instead of having to work with graphics libraries and low-level code, you just put a lot of what the user will deal with into Java or bung most of it into a browser-based application. Compared with something like an iPhone or Palm Pre, the Android interface is still a bit clunky but there is a reasonable prospect that software writers will improve that as Android phones become more common.

You've got to like Mentor CEO Wally Rhines' style here: while Wind River is saddled with the job of pushing Intel-approved Moblin - which hardly anybody has heard of - into the appliance business, Mentor can employ a software environment that a lot more engineers will have encountered. OK, Android isn't in many handsets but it's still in front of way more people than Moblin. Something I have argued in favour of for a number of years is that what most embedded applications really need is a half-decent scripting environment that lets teams prototype the user interface while leaving the core functions untouched. This lets companies contract out the industrial design, which can include the user interface, while the important bit that does the job - such as controlling a wash cycle - remains within the core team.

There is one small problem with Mentor's plan at this stage. Embedded Alley has, so far, focused on MIPS processors rather than ARM or Intel architectures. I've nothing against MIPS but most of the action is going to take place around the other two simply because of the momentum that the mobile handset business has. The handset business provides the volume needed to make complex multicore processors cheap and it's relatively simple to divert those devices into other markets. This is the plan that Texas Instruments has followed with its OMAP devices. You have to know the secret mobile-maker's handshake to buy some OMAPs but others have been repackaged for the wider embedded business.

However, I would expect Mentor to place more emphasis on ARM and Intel once Embedded Alley is part of the fold.

Synopsys has joined ARM, IBM, Chartered Semiconductor and Samsung in a programme intended to get the 32nm Common Platform off to a running start. The EDA company will use the Lynx design environment that it built in the past few years to hook together a ready-made flow that will act as the gateway to production at the Common Platform foundries. ARM, naturally, will supply libraries and IP cores that have been tried out on test chips.

By providing a ready-made environment, the aim is to pull customers into 32nm design much more quickly than they might have without having this kind of support, according to Kevin Meyer, vice president of marketing at Chartered. "We believe customers can come to this technology a full year earlier," he claimed.

IBM is placing a bet that 32nm will provide the breakthrough for Common Platform. The shift to a high-k, metal-gate transistor may provide the technical advantage that Common Platform needs over TSMC, the current leader in foundry production. TSMC plans to bring in a 28nm process with a high-k, metal-gate but IBM has a little more experience in that department having worked on it for an internal 45nm process, although that never saw the light of day as a Common Platform-supported process.

The 32nm node could prove to be an important one, much like 130nm or 65nm are today. Mark Ireland, vice president at IBM Semiconductor Platforms, said companies are increasingly skipping nodes.

The conclusion is that they are more likely to skip from 65nm to 32nm for a density increase rather than to 45nm or 40nm. There are some doubts as to how easy it will be to move to 22nm if some techniques to produce accurate images on the chips do not work.

"From what we have seen, 32nm will be a very long lived node. There is a lot of excitement about it and a lot of companies will be skipping to 32nm," Ireland claimed.

The rise of chip design in Asia is playing a large role in the decision to roll out this kind of environment. Meyer pointed to the large number of 65nm designs being done now in the Far East. China is picking up steam in this department, partly helped by attractive subsidies for companies to perform advanced designs in that country. This has encouraged companies that used to buy off-the-shelf chips to do their own and push towards the advanced nodes to get the per-die price down.

Normally, EDA companies hang back from doing this kind of advanced work. It's generally better to see the colour of the customer's money before ploughing in a lot of development. John Chilton, senior vice president of marketing at Synopsys refused to discuss the commercial arrangement that brought the company into the Common Platform fold. But it is interesting that there is only one EDA company there - TSMC has been very careful to maintain neutrality when dealing with the major EDA companies when putting together its reference flows.

Driving demand is very important for the foundries - this deal provides a way of easing access to what will be a tricky process to design for. But the commercial arrangements will be very sensitive. ARM is in a pitched battle with Intel to provide the dominant architecture for mobile Internet devices, exactly the target for the 32nm and 28nm Common Platform processes. The Common Platform axis has aligned itself around ARM and its IP - TSMC is the foundry of choice for Intel Atoms. But Synopsys is the 'official' primary tool supplier to Intel as well as for this programme. Or was that the former primary tool supplier to Intel?

In the latest issue of IEEE Spectrum, Professor Clayton Christensen (of The Innovator's Dilemma fame) and colleagues describe their role in trying to improve efficiency at a big chipmaker, taking much of their inspiration from the Toyota Production System. The company goes unnamed, but I assume that it's not Intel as some of what they describe was put into action at the number-one chipmaker in its Copy Exact programme.

One thing that the authors note - and one that strikes just about anybody who visits a fab - is how much manual intervention chipmaking processes need. They are more automated than they used to be. Intel was able to turn two fabs into one at Leixlip in Ireland thanks to a monorail for wafers that runs between corridors of equipment. It was not long ago that boxes of wafers travelled on trolleys pushed by people.

There were good reasons for not automating heavily. It was easier to train people to deal with small changes and a given piece of equipment can take hours to process a full box of wafers. Why go to the bother of automating when people can react in plenty of time?

But, as Christensen and colleagues write, having a lot of on-the-spot manual changes makes it difficult to trace problems back to their source. By having the processes formalised and recorded, and then working on ways to improve them, the authors claim they were able to reduce wafer processing costs by 12 per cent and cycle time - perhaps more important in an environment where being able to respond to sudden changes in demand has become critical - by 67 per cent.

Details will differ, but the project conducted by Christensen and colleagues sounds very similar in concept to the Next Generation Fab proposals coming from the Japanese fab owners and from work at Sematech. In the past, operational efficiencies came from Moore's Law and from increases in wafer size. But the incremental costs are now so high that it makes sense to look more at the minutiae of operating fabs. It is perhaps a further symptom of the way in which Moore's Law is pricing itself out of the mainstream.

The piece ends on a slightly bizarre note - one so optimistic you start to wonder whether the authors have been anywhere near chip design. They argue that, for manufacturers to make the most of the efficiencies in applying the Toyota system to their operations, they need to remake the design process. The answer is to make things modular. Yes, slap that forehead hard. If only we realised that was the answer!

This is the point at which Christensen and colleagues need to do an analysis of what costs money in design. Little things like embedded software. It's true that chipmakers could do more to standardise and cut costs but the idea of plug-and-play components that let you slap together a system like a Dell PC is only workable in certain environments. And, few of those environments demand full-custom chips - microcontrollers and field-programmable gate arrays (FPGAs) will often do just fine.