In R2018a, Simulink® and Embedded Coder® support a variety of mechanisms to minimize unnecessary data copies in generated code. This produces more efficient code in both speed and memory consumption.
In addition to automatic buffer reuse and user-specified reuse through storage classes, R2018 introduces an intuitive option for engineers to suggest data reuse through the use of signal labels. This approach makes it easy to annotate models, does not force buffers to be global, and allows further code optimizations and reuse.
This presentation is from Coder Summit Talks, which are live recordings of developers and engineers debuting their best MATLAB® and Simulink Coder™ optimizations and examples at an annual technical interchange.