MathWorks Optimizes Code Generation Across the Entire ARM Cortex Family
October 31, 2014 by John Day
Extends benefits of Model-Based Design for embedded developers by generating optimized code for Cortex-A processor family
Embedded systems developers using ARM Cortex processors in the automotive, aerospace, signal processing and medical device industries can now automatically generate code, speeding execution time and reducing memory usage. This ability to generate optimized code helps small design teams extend Model-Based Design across applications and devices including motor control, power electronics, audio, and communications.
Building on CMSIS library support for Cortex-R and Cortex-M cores, Embedded Coder now offers Ne10 DSP library support, which helps generate optimized DSP filter code for the full line of ARM Cortex-A processor-based devices.
Simulink also offers built-in target support for ARM-based hardware devices including BeagleBone Black, Xilinx Zynq, STMicroelectronics Discovery Board, and Freescale Freedom Board. Third-party integrations exist for other popular ARM-based devices including Texas Instruments Hercules and Analog Devices CMx40 hardware.
Model-Based Design offers engineers three key benefits throughout the software and hardware design process:
• By using Simulink for desktop simulation of systems, engineers can catch errors early and optimize designs.
• Automatic code generation with Embedded Coder can help to prevent manual coding errors and optimize code.
• Automated testing of model code using processor-in-the-loop with Embedded Coder saves time via test reuse and improves code quality.
“Our partnership with MathWorks continues to help developers efficiently produce accurate code and their support for our full line of Cortex processor-based devices is another important step,” said Richard York, vice president, embedded segment, ARM.
“With Ne10 optimized code, embedded systems developers working with Cortex-A processors can now realize the cost and time-to-market benefits of Model-Based Design with fast ARM NEON™ code performance on ARM cores.”
“Increasing interest in image processing, radar, and communications is driving more use of ARM Cortex-A-based devices for a wide range of consumer and commercial devices,” said Paul Barnard, design automation marketing director, MathWorks.
“Consequently, embedded software engineers and system level designers often need to simulate their software and systems within the same environment. The ability to generate optimized code will serve as an on-ramp to Model-Based Design for groups of engineers developing ARM-based devices, by allowing them to start with modeling while knowing they can generate code for their embedded system.”