By John Keller
Editor in Chief
Competition for microprocessor preference in the aerospace and defense embedded computing industry is heating up with the introduction of a next-generation Power Architecture processor by Freescale Semiconductor Inc. in Austin, Texas, as an alternative to the 2nd Generation Core i7 processor by Intel Corp. in Santa Clara, Calif.
Freescale’s new QorIQ Advanced Multiprocessing (AMP) series brings back floating-point processing that had been standard in previous generations of Freescale AltiVec processors popular for aerospace and defense digital signal processing (DSP) applications, such as radar, sonar, electronic warfare, and signals intelligence. Freescale had abandoned the AltiVec floating-point engine in its QorIQ microprocessors a few years ago because of a perceived lack of interest from core target markets, which served to alienate many of the company’s aerospace and defense customers, who had come to depend on the AltiVec architecture as one way to avoid using dedicated DSP chips in signal processing-intense applications.
Intel Corp. in early 2010 introduced its Core i7 microprocessor architecture, which offered floating-point capability largely for graphics-intensive personal computer applications. Aerospace and defense embedded computer manufacturers quickly realized the Core i7 helped fill the void left by the discontinued AltiVec, and jumped onto the Intel developmental roadmap.
Now Freescale is back, with Alti- Vec and floating-point processing. The big question is whether Free- scale’s renewed support for the kind of floating-point processing that aerospace and defense embedded computing needs is too little, too late.
The big military and aerospace embedded computing suppliers, including Mercury Computer Systems in Chelmsford, Mass., Curtiss-Wright Controls Embedded Computing in Ashburn, Va., GE Intelligent Platforms in Charlottesville, Va., and Extreme Engineering Solutions Inc. (X-ES) in Middleton, Wis., are heavily supporting the Intel Core i7 in their products. Will the embedded computing heavy hitters support the latest Intel and Freescale microprocessors in equal measure? The market has not yet given us a clear indication.
For now, all the major aerospace and defense embedded computing manufacturers are expected to offer products based on the QorIQ AMP processor, which will have four times the processing performance of the previous generation of Freescale processors, says Glenn Beck, Freescale’s segment marketing manager for aerospace and defense.
Each core in the multi-core QorIQ AMP has an AltiVec processing engine, and each core is dual-threaded. The chip is based on 28-nanometer technology, as opposed to 45-nanometer technology for previous versions of Freescale’s microprocessors. The first microprocessor in the QorIQ AMP family will be the T4240, which should sample in early 2012.
The new processors are designed to support heavy network traffic. “Network traffic is increasing dramatically—voice, data, and video,” Beck says. “Most military air platforms are moving to being unmanned, and from generation one to generation two of these platforms, the amount of sensing capability has exploded in the amount of data to be processed. Military users have a need for autonomous decision making at the point where the data is generated.”
Freescale engineers are certifying the QorIQ AMP processor with the National Security Agency (NSA) for resistance to theft of data, theft of intellectual property, and attempts to counterfeit the device, Beck says.
It remains to be seen how attractive the new QorIQ AMP processor will be in aerospace and defense applications. What we do know is there’s a new chip in town.