CPU Code Optimization with ARM Cortex-A CMSIS CRL and Neon SIMD

Embedded Coder^® Support Package for Qualcomm^® Hexagon^® Processors provides a code replacement library (CRL), ARM Cortex-A CMSIS, to generate calls to the CMSIS-DSP library optimized for ARM^® Cortex^®-A processors. This CRL is supported when you work with CPUs - Qualcomm Android or Qualcomm Linux.

To use ARM Cortex-A CMSIS CRL, go to Code Generation > Interface, click Select corresponding to Code replacement libraries, and then select ARM Cortex-A CMSIS and move it to Selected code replacement libraries - prioritized pane.

For details about the CMSIS code replacements for the various operations that you implement on the CPU, refer to:

• Supported MATLAB Functions with CMSIS Library for ARM Cortex-A Processors

• Supported Simulink Blocks with CMSIS Library for ARM Cortex-A Processors

• Supported DSP System Toolbox Blocks with CMSIS Library for ARM Cortex-A Processors

• Supported DSP System Toolbox System Objects with CMSIS Library for ARM Cortex-A Processors

The support package also supports Neon v7 instruction set to generate SIMD (single instruction, multiple data) code for Qualcomm Android or Qualcomm Linux CPUs . To enable this, go to Code Generation > Optimization, and select Neon v7 for the Leverage target hardware instruction set extensions parameter.

Optionally select Optimize reductions to generate SIMD code for reduction operation loops, and select FMA to enable support for the fused multiply-add (FMA) extension. For more information, see Optimize reductions and Generate SIMD Code from Simulink Blocks for ARM Platforms.