ARM has also substantially expanded its processor core design sockets -- both to high-end and low-end markets. On one hand, the company has bravely marched into the multi-core future, and has relentlessly rolled out a series of Mali graphics processor units and video engines.

On the other hand, ARM this year made a genuinely strong foray into the microcontroller market. ARM took the first step by offering initial seed funding to Luminary Micro. Then earlier this year, Texas Instruments -- by acquiring Luminary Micro -- turned ARM into a household name in the microcontroller market.

Ninety ARM processors are shipped every second -- more than any other 32-bit processor IP supplier, the company's Web site claims.

So far, everything you hear about ARM sounds peachy. You're impressed.

But then, ask yourself: If you were running ARM's processor division, what's your biggest worry right now?

The answer is "plenty."

Not least among ARM's headaches is the serious competition posed by a gargantuan microprocessor company (Intel). You might lose sleep over how best to help your licensees navigate uncharted shoals of multi-core and multi-threaded CPUs. You might also be concerned about whether graphics companies like Nvidia, AMD or Imagination will become serious rivals as you move forward with your Mali GPU.

Most of all, you worry that you could be overstretching yourself.

During an interview at the ARM European Technical Conference here this week, Mike Inglis, executive vice president and general manager of the processor div. at ARM, said, "As ARM gets deployed in a much broader market, the industry's demand on us has dramatically increased." He said, "I worry if we are moving fast enough."

Inglis is also mindful of critical responsibilities in building software eco-systems, a must for the future growth of ARM. But he also noted, "We need to make sure that we are not duplicating efforts."

As befits any executive in the electronics industry today, Inglis is worried about the economy. "We are beginning to see a recovery in this industry," he said, "but nobody knows if consumer spending will come back."

During this Great Recession, Inglis said that ARM accelerated the company's efforts particularly in engineering. This year, ARM has also kept a keen eye out for "strategic investment opportunities." ARM and Qualcomm, for example, recently led a group of investors to help network-on-chip IP supplier Arteris Inc. raise $9.7 million.

Inglis said ARM has made similar investments in a series of companies -- although not publicly announced -- in 2009. ARM's efforts have been critical, especially during this year's drought in VC money. Inglis said, "We made commitments to some start-ups, hoping that this will help them attract money from other sources." ARM's original funding in Luminary Micro is an example of how ARM's seed investment helped not just the start-up, but ARM and ARM's partners.

As widely successful as ARM has been on the mobile handset market, what's remarkable about the company is that it has never rested on its laurels.

In the third quarter of this year alone, the company claimed that 15 new processor licenses signed, including four next-generation processors, for a broad range of markets including digital TV, microcontrollers, hard disk drives and networking applications

During the interview, Inglis stated his ambition bluntly: "We want every socket MIPS is in today."

ARM's invasion into new market areas beyond mobile is already gaining a solid ground. Getting STMicroelectronics, one of the world's leading set-top-box chip makers, to adopt the ARM Corte-A9 MP Core processor, in addition to the Mali-400 graphics processor, for its upcoming set-top-box and digital TV system-on-chip ICs, was a huge win for ARM. ST, traditionally, has depended on home-grown processors to drive its digital consumer products.

Inglis acknowledged that ARM's ST persuasion has taken "seven years."

Similarly, ARM recently announced that it has talked Infineon Technologies AG into using ARM in security controllers for chip card and security applications.

Infineon has become the only ARM partner with an ARM architecture license specifically for security applications. Under this agreement, Infineon is able to integrate its innovative security measures into the heart of the CPU core implementation, while maintaining compatibility with the standard ARM processor instruction set.

ARM hasn't relaxed efforts to win as many sockets as possible in the netbook, smartbook and standard computing market, either. Inglis said, "This will take five to 10 years. We know software is the key, and a quiet battle is brewing inside the software community."

New compilers?

Many SoC designers -- and ARM licensees -- are struggling on one issue in today's environment of multiple GPUs and MPUs: Is there a comipler that can automatically spread load energy efficiently across those multiple cores in an SoC?

ARM's answer is, no, not yet.

Acknowledging that "huge software challenges are coming," Inglis said, "Many SoC vendors are, just about now, running into an 'oops' phase." He warned that the industry is still in a very early stage of using multiple cores for their SoCs. "Real SoCs based on multiple cores to run real applications are just about to come out now. They are being debugged. Many engineers are suddenly asking, 'how do you run all these tasks in so many cores?'"

ARM has "profiling tools" that let designers look at software and see where the bubbles (energy peaks, for example) are, explained Inglis. But there are no tools that could actually help them design an SoC by spreading the load among different cores in the most energy efficient way. "Right now, many designers are using their existing codes to run on multiple cores and re-profiling them."

While there is no silver bullet yet, Inglis acknowledged that the industry has seen a considerable interest in OpenCL, especially after Apple's investment. OpenCL could be leveraged for designers to get a whole system's view. Or it may be applicable to ARM's GPU, or Neon technology.

"There are a lot of possibilities but much of our development work is still in our labs," said Lionel Saives, senior field applications engineers at ARM. "OpenCL provides parallelism but it tends to work better in number crunching, scientific applications."

How do you compete with Intel?

Nobody should take Intel lightly, despite Intel's numerous failed attempts to crack the embedded market, including mobile phones and other consumer devices.

Inglis believes two factors will work for ARM. One is a cultural difference between the two companies' approaches toward the embedded market. Another is the business model.

Historically speaking, Intel has always let performance improvement drive the company's roadmap. In contrast, ARM has emphasized energy efficiency in reaching out to broader-based embedded applications. Further, Inglis said, Intel's roadmap stops at the $6 processor.

But at ARM, "we are quite happy to go cheap -- well below $6," in search of microcontroller apps.

But the second and possibly bigger difference between ARM and Intel lies in the business model, according to Inglis. "With ARM, 600 companies are using ARM architecture to develop innovative processor solutions.

For Intel, X86 is offered by one or two players." Despite a huge reduction of processor IP providers in this market, "Engineers always love choices. The business needs choices," said Inglis.

And the engineering community is about to see a whole new set of choices unfold, stressed Inglis, including: "GlobalFoundries vs. TSMC; Google vs. Microsoft; and ARM vs. Intel."

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved