By The EEFocus Staff

FPGAs are booming in China. When Clement Cheung, director of marketing and applications at Xilinx Asia Pacific showed up to give a speech recently, he was worried not many people would come. He need not have worried.

The sales of all the major FPGA vendors show a significant bump in sales to the Asia/Pacific region. In Xilinx’s case, they exceed sales in North America, and there has been a huge increase in the number of FPGA engineers and engineers using FPGAs inside of China. This follows the shift of manufacturing to the Asia/Pacific region earlier this decade.

But Cheung said China is particularly important to FPGA vendors: “We pay more attention to the growth of Chinese enterprises because only they can truly influence the country.”

The proportion of FPGAs in communications and in the enterprise fell sharply following the 2001 downturn, but FPGA vendors managed to weather the latest downturn relatively intact by limiting their presence in communications to less than 50% and by hedging across multiple other markets.

“The main reason [for growth] is the 3G network deployment in countries like China,” said Clement Cheung. This was also one of the key reasons that the Asia/Pacific reason posted strong growth. Xilinx’s growth in China has been in the double digits, Cheung said.

FPGAs have been particularly popular because they lower the barrier of entry for design companies. It currently costs millions of dollars for an ASIC mask, but an FPGA is a much less expensive alternative. Huawei, which applied for the most patents inside of China in 2008, based a lot of its work on FPGAs.

“In China, FPGA engineers have a larger number and all of them stand on the same starting line,” Cheung said. “But FPGAs are more than a pure chip game. At present, most engineers need to consider signal integrity, layout, timing and other system-level issues.”

Xilinx CEO Moshe Gavrielov pointed out that the device functions had changed in the FPGA industry over the past 25 years, evolving from the simple circuit such as peripheral interface and glue logic to the main chip of whole system. “The concept of platform was not obvious because customers didn’t need too many application designs in the past. Today, the chip, software and whole design environment all need to be combined together for the project design.”

–EEFocus is the Chinese media partner of System-Level Design.

By Wang Xiaodan

Strength, vision and perseverance are a must for all the enterprises that have successfully carried out university programs. That ranges from equipment donations, university competitions and joint laboratories all the way to teaching material support, R&D funds, and students’ practical training. And all of this requires not only money, but also the integration of resources of the enterprise itself for target institutions, industrial chain partners and government departments.

In this sense, university programs have become a long-term strategy for enterprises. There are hardly any results for short-term operations, performance is difficult to assess, and a lot of enterprise resources are involved.

Interestingly, those enterprises that have their own “research institutes” are loyal practitioners of university programs. In general, internal research institutes have projects that last for three to seven years, allowing them to carefully plan their university programs and develop the deep connection necessary between the university programs and their own future. These enterprises include Intel, IBM, Microsoft and Google, among others.

The reason is simple: They have established their present position, and are willing to lay out the future. Just consider that an entire generation of young Chinese students may become your potential customers, potential buyers, potential partners and even the future policy-makers. And even without any direct business connections, they will have favorable impressions on your brands.

Intel is a good example of how a company is building loyalty in China, and its approach involves several steps. The first includes teachers. Through course collaboration and teacher training, Intel has established cooperation with more than 100 universities. Take the multicore university program as an example. Intel first invited teachers from five domestic universities to develop courses and opened its first teacher training class in 2006. So far, the company has held 11 teacher training classes, and trained 400 front-line teachers, who have cultivated nearly 20,000 students afterwards. Three courses on the multicore technology have become state-level high-quality courses.

The second step involves students in the form of student competitions, scholarships and enterprise internships. For example, Intel sponsors the National Embedded System Design Competition and the National Multicore Technology Programming Competition, introduced the “Intel Scholarship” program, and cooperated with the Ministry of Education in the founding of Intel Internship Base. This base, located in Intel Asia-Pacific R&D Center, Zizhu Science-Based Industrial Park of Shanghai, has received 600 students for exercitation. All students are assigned with mentors, and participate in the R&D process just as full-time staff.

Meanwhile, in order to coordinate with Intel’s deep embedded strategy, one general-purpose university textbook based on the examples of the x86 architecture and Atom processor will be promoted in Chinese educational institutions next year. This is how Intel intends to weaken ARM’s grip on the Chinese electronics market. Presently, Chinese students majoring in electrical engineering learn and use ARM processors and tools, and regard their embedded design skills based on ARM’s architecture as a stepping stone for the job interview. In this regard, Julia Zhu, Education Manager of Intel China Ltd, said, as long as Atom’s technology and platform are open, Intel will still have winning chance in the competition.

“Teachers are familiar with the x86 architecture, and students have been learning the Principles of Microcomputers based on x86 architecture. The problem is that this textbook is outdated,” she said, “Therefore, by leveraging Atom’s new technology and platform we can help institutions to complete the curriculum reform and continuously provide teachers with better and newer technology and equipment support. In this case, we also have advantages our competitors do not have.”

Jun Ge, executive director of Intel China Ltd. in charge of external cooperation, noted that Intel has a China Strategy Office that specializes in what products Intel should launch in China in 10 to 15 years. With such a long-term outlook, Intel certainly needs to cultivate its Chinese fans over the next 10 to 15 years, which is the goal of Intel’s University Program Planning Department.

EEFocus is the Chinese media partner of System-Level Design.

By Ed Sperling

Doing business in China is more complicated and frustrating than it looks. The U.S. Information Technology Office testified before the World Trade Organization that there has been some progress—and some roadblocks—in recent years.

And despite the EDA industry’s lackluster first half, chip sales are up 5% month on month, according to the Semiconductor Industry Association. At what node?

STMicro apparently is quite happy with Synopsys’ USB 2.0 and Ethernet IP. The company achieved first-pass silicon for its STM32 SoCs. Yes, folks, it can be done (sometimes).

Cadence extended its TLM hardware/software co-verification to allow software developers to verify and debug their software earlier in the design process. Incisive now has debug support for C/C++ compilers from Virage Logic’s ARC, ARM, GNU and Green Hills Software.

ARM and GlobalFoundries signed a strategic partnership for an SoC enablement program built around ARM’s IP and Globalfoundries’ high k/metal gate manufacturing capabilities.

Intel inked a deal with the government of Macedonia to improve the quality of education by accelerating e-learning. Considering the last major figure to emerge from this region was Alexander the Great, it may be hard to underestimate the effects of this one.

More designs by fewer companies puts renewed value in the EDA world, according to Ray Bingham, managing director of General Atlantic, a private equity firm. Bingham, the former CEO of Cadence Design Systems, talks about what is changing in technology and globalization.

One of the subtle differences hidden in global design data is exactly when design starts peak in various regions around the globe. While the total number of designs remains relatively consistent, the activity in each area does not.

In Asia, for example, the biggest months are at the beginning of the year and between May and July. In Europe, the numbers are highest at the beginning of the year, with moderate activity in April, May and June. And in North America, design activity typically starts strong at the beginning of the year and peaks in May, July and September, just in time for fourth-quarter delivery of products.

While this is likely to become skewed in 2009 because of the downturn, there are a couple of interesting trends at work here that persist from year to year in normal growth years. First of all, much of the design work in North America and Europe is done at similar times and is of similar types, except that European workers typically take their vacations during the summer months. In North America, design continues even if some of the team is on vacation.

In Asia, many of the designs are tied into local markets, which run on a different schedule. The biggest retail months in China are in November, during the national holiday, and during the Lunar New Year. Those designs aimed at export markets typically have been less technologically advanced or narrower in focus than many designs out of Europe, North America and Japan.

–Ed Sperling

Offshoring started out as a less-expensive way of developing enterprise applications, but outsourced software development is beginning to move much deeper into the system-level design world.

So far, development has evolved from just productivity applications to embedded software. Even large-scale system design is being outsourced. The next step is to outsource pieces of system-level designs, something that could have a dramatic impact on the roles of engineers in the United States, Europe and Japan.

System-Level Design caught up with Gordon Brooks, CEO of Symphony Services, a 5-1/2 year old company based in Palo Alto, Calif., to talk about what’s changing in this world.

Gordon Brooks

SLD: What’s your focus at the moment?

Gordon Brooks: We started working with independent software vendors building commercial software products. That was our core. We also do commercial-grade software development for non-ISVs. The third piece is embedded software engineering. It’s the software for automobiles or consumer electronics or telecom or storage devices.

The embedded focus is particularly interesting for system-level design engineers, because they’re being asked by their customers to provide more than just the hardware.

Our clients range from Motorola to Hitachi and small and midsize companies. We create robust applications for these devices, from the chip up into the application. These devices now have to run navigation, video and all the video systems. There are companies that outsource embedded software, but most go the other way—from the chip down into the hardware and out to manufacturing. We focus on going from the chip up into robust applications.

This isn’t a place where most chip developers have expertise, right?

No, and they don’t think it’s core, either. It’s not that it isn’t core to their business, but they don’t feel like they have to own it in-house.

What’s the reaction inside the chip companies?

These decisions are made pretty high up by the CEO, CFO or business president—it depends on the size of the deal and what you’re doing. The reaction is mixed. There’s the obvious threat of companies outsourcing. But it is a big trend. We just signed a contract to do wingspan structures for Airbus. We’re working with the subcontractors, who do the manufacturing, and they’re outsourcing the designs to us.

What can go wrong in these projects?

First of all, there has to be labor arbitrage. But beyond that, there often are quality or scheduling problems or access to talent. It isn’t just about cost anymore. The biggest failure is when people outsource at the task level instead of a subset or goal level. If someone defines task and whoever they outsource to doesn’t have the big picture and the context, then you have a lot of people asking questions. The people who have been successful is when you take a define-able component and outsource it end-to-end. That’s where you get innovation.

Is it more difficult when pieces are being outsourced than doing pieces internally?

It’s no different than if you do something in San Antonio, Texas, and another piece in Helsinki, Finland, even if you do it all yourself. It’s how you break the work up. If you break it into too many pieces, or if you don’t provide the ability to deliver a component of some kind, then it makes it hard to do.

Where are you finding the best talent for this kind of development?

India is 10x above everyone else from a capability and background standpoint. China is coming up, and it has some advantages because of the manufacturing. But there’s a lot of manufacturing going to India, too. There are niche areas in other places, but at scale the only two places are China and India.

What are the savings from a cost standpoint?

When you consider everything, including communications, it’s still less than half the cost of developing it in the United States.

Is that changing?

Yes, but it takes a long time for equalization to happen. The wage increases that are being reported are also due to promotions and additional skills. Less than half of the reported increases are really due to wage inflation. Quite frankly, in India this year the average increase will be less than 5 percent.

How about design outsourcing?

We’re beginning to get into that. Coupling that with what’s happening at the chip level and the architectural level is the next step.

Is that skill level available in India and China?

Yes. These countries are designing military planes and weapons. There’s a very strong capability there.

How about designing chips?

We haven’t started on that yet. It’s a little early, but it will happen. People are getting used to getting everything around the chip outsourced first.