Use Target Hardware Instruction Set Extensions to Generate SIMD Code from Simulink Blocks for ARM Cortex-A Processors
Note
This workflow requires an Embedded Coder® license.
This topic shows how to generate SIMD code from Simulink® blocks using the model configuration parameter Leverage target
hardware instruction set extensions and the Embedded Coder Support Package for ARM®
Cortex®-A Processors. You can also generate SIMD code by using the code replacement library
ARM Cortex-A
, as described in Optimize Code for ARM Cortex-A Processors (Embedded Coder). The code replacement library also applies
other optimizations, whereas the configuration parameter applies only SIMD
optimizations.
Open the Embedded Coder app from the Apps gallery in the Simulink model toolstrip.
In the C Code tab that opens, click Settings > Hardware Implementation.
Set the Device vendor (Simulink) parameter to
ARM Compatible
.Set the Device type (Simulink) parameter to
ARM Cortex-A (32-bit)
orARM Cortex-A (64-bit)
.On the Code Generation > Optimization pane, for the Leverage target hardware instruction set extensions (Simulink Coder) parameter, select the instruction set extension supported by your processor. For this example, select
Neon v7
.Optionally, select the Optimize reductions (Simulink Coder) parameter to generate SIMD code for reduction operations or the FMA (Simulink Coder) parameter to generate SIMD code for fused multiply add operations.
On the Code Generation > Interface pane, under Software environment, set Code replacement libraries to
None
, and clear Support non-finite numbers.Depending on the blocks you have used in your model, you might have to change additional block-level settings to generate SIMD code using this workflow. For more information on these settings, see the Extended Capabilities > C/C++ Code Generation sections in the respective block reference pages.
Generate code from the model.
The SIMD instructions are the intrinsic functions that start with the identifier
v
in the generated code. These functions process multiple data in
a single iteration of the loop because the loop increments by four for single data types
and by two for double data types. For models that process more data and are
computationally more intensive, the SIMD instructions can significantly speed up the
code execution time.