NXP decides to make the Cortex switch

The number of microcontrollers shipped has only just "crawled past the 100,000 level" according to ARM CEO Warren East. But the company has signed its eighth licence for the Cortex-M3, bringing ARM7 stalwart NXP Semiconductors into the fold.

NXP Semiconductors has decided to embrace ARM's Cortex-M3 as its favoured architecture for 32bit microcontrollers, joining STMicroelectronics and others in adopting the processor core, which several years after its initial launch has only just passed the 100,000 mark in terms of unit shipments.

Up to now, NXP used the older ARM7 architecture as the basis for its 32bit microcontrollers. The Cortex-M3 was, unlike the ARM7, designed primarily as a microcontroller core rather than a general-purpose processor. However, it is not binary compatible with the ARM7 as it only runs Thumb instructions rather than Thumb and the full 32bit ARM instructions. The compatibility issue was a factor in NXP's delay in deciding whether to go to the M3.

"The beauty of delaying a little is that the toolchain is now stable," said Geoff Lees, vice president and general manager of the microcontroller product line at NXP. "Our ARM7 family has been a runaway success. There is a huge demand for ARM cores: I don't see that much of a difference between the specific cores, just a general demand for ARM," Lees added, noting that it will take a little time for the M3, as a core, to catch up with the larger parts in NXP's range of ARM7-based parts.

Lees said the most successful families in the ARM7 range are those with the higher-end peripherals such as USB interfaces and display controllers. "The biggest trend we are seeing is customers going straight in with large code sizes and using 512KB flash parts in volume. That was non-intuitive to us: we thought the smallest memory would have the highest volume. The 32bit space seems to be encouraging larger code sizes."

Because the ARM7-based parts are likely to have more capable peripherals on them than the initial M3 designs, NXP expects to continue with the existing architecture in parallel. "The real-time advantages of the M3 are not top of the list for many of those applications. So, with M3, is there a benefit in terms of code size? Is it worth transferring over from ARM7 just yet? Once the customer decides which is the majority infrastructure, they will go that way. But I think we will continue to offer an overlap in the meantime."

Warren East, CEO of ARM told analysts that initial progress with the M3 had been slow and that shipments had just "crawled past the 100,000 level" but that momentum was picking up.

The first company to launch M3-based parts was startup Luminary Micro, entering the market in spring 2006. The M3 now has eight licensees.

NXP's deal with ARM gives the chipmaker access to all of the Cortex family processors, including the recently announced Cortex-A9 MPCore multicore processor and the ARM Mali family of graphics processing units (GPUs).

Lees said the first M3-based parts will be out later in the year, although the company will show early versions to lead customers before then. The company will continue to use the Memory Accelerator Module (MAM), initially used by the company's ARM7-based parts, to improve access times with embedded flash memory.

Image: NXP will launch the first of its M3-based microcontrollers in 2008

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them