Performance

Guidelines for improving generated code performance

When designing MATLAB^® algorithms intended for code generation, observe the following guidelines to improve the performance of generated code.

Blocks

Include MATLAB code in Simulink models

`coder.areUnboundedVariableSizedArraysSupported`	Check if current configuration settings allow unbounded variable-size arrays (Since R2024a)
`coder.inline`	Control inlining of current function in generated code
`coder.inlineCall`	Inline called function in generated code (Since R2024a)
`coder.nonInlineCall`	Prevent inlining of called function in generated code (Since R2024a)
`coder.unroll`	Unroll `for`-loop by making a copy of the loop body for each loop iteration
`coder.const`	Fold expressions into constants in generated code

Unroll for-Loops and parfor-Loops
Control loop unrolling.
Inline Code
Inlining eliminates the overhead of a function call by replacing the function call with the body of the function.
Avoid Data Copies of Function Inputs in Generated Code
Generate code that passes input arguments by reference.
Generate Reusable Code
To reuse functions, save them in files or integrate external functions into generated code.
Control Run-Time Checks
Improve performance by disabling run-time checks.
LAPACK Calls for Linear Algebra in a MATLAB Function Block
LAPACK function calls improve the simulation speed of MATLAB Function block algorithms that call certain linear algebra functions.
BLAS Calls for Matrix Operations in a MATLAB Function Block
BLAS function calls improve the simulation speed of MATLAB Function block algorithms that call certain low-level vector and matrix functions.
FFTW calls for fast Fourier transform functions in a MATLAB Function Block
FFTW function calls improve the simulation speed of MATLAB Function block algorithms that call MATLAB fast Fourier transform (FFT) functions.