Main Content

Operating Point Behavior

Use a model operating point with a block, S-functions and model changes.

Change the States of a Block Within Operating Point

  • Use the loggedStates property of the Simulink.op.ModelOperatingPoint object to get or set the states of blocks. If xout is the state log that Simulink® exports to the workspace, then the loggedStates field has the same structure as xout.signals.

  • With dataset logging, saving the model operating point captures the current logged states snapshot as saving the final states without the operating point would, and stores it on the ModelOperatingPoint object. When you query this logged state, a copy of the saved dataset is returned.

  • With dataset logging format, the answers returned by ModelOperatingPoint always match those returned by state logging (non-Operating Point save). Logged states for blocks inside of accelerator-mode model reference are visible and editable. Logged states for many blocks with custom operating point functions, for example, Simscape, model reference accel and variable transport delay are editable.

  • You cannot change the states that are not logged. Simulink does not allow this modification as it could cause the state to be inconsistent with the simulation.


You can use APIs for C-MEX and Level-2 MATLAB® S-functions to enable S-functions to leverage the model operating point. For information on how to implement this functionality in an S-function, see S-Function Compliance with the ModelOperatingPoint.

S-functions that have PWork vectors, which store pointers to data structures but do not declare their operating point compliance level or declare it to be unknown or disallowed do not support operating point. For more information, see S-Function Compliance with the ModelOperatingPoint.

Model Changes and Operating Point Restore

After saving the operating point of a model, you can change the model and restore an operating point with those changes.

  • You can rename a model between saving and restoring its operating point.

  • The operating point interface checksum is primarily based on the configuration settings in a model and sample times used in the model. You can make some nonstructural changes to the model between saving and restoring an operating point. In the Configuration Parameters dialog box, in the Diagnostics pane, use the Operating point interface checksum mismatch diagnostic to track such changes. You can then find out if the interface checksum of the restored operating point matches the current interface checksum. See Operating point restore interface checksum mismatch.

    You can make the following nonstructural changes to the model without affecting your ability to restore a previously saved operating point:


    These modifications may change the number of sample times in the model. This can cause the model's interface checksum to be different from the operating point save and restore checksum. Configure the Operating point interface checksum mismatch diagnostic to display a warning (default), error, or none to not compare the checksums.

  • You cannot make structural changes to the model between the time you save the operating point and the time you restore the simulation using the operating point. Examples include, adding or removing a block after saving the operating point, changing the sample time of a model, and changing the type of solver from variable-step to fixed-step.

  • Mismatches can occur when you try to simulate using a solver that is different from the one that generated the saved operating point. Simulink permits solver changes. For example, you can use the ode15s solver to solve the initial stiff portion of a simulation and save the final operating point. You can then continue the simulation with the restored operating point using ode45. In other words, this diagnostic helps you see the solver changes but does not signal a problem with the simulation.


When you use a variable-step solver with the maximum step size set to auto, Simulink uses the maximum step size from the restored ModelOperatingPoint object for the new simulation. To ensure that the concatenated operating point trajectory of two simulations matches that of an uninterrupted simulation, specify a value for the maximum step size.

Limitations of Saving and Restoring Operating Point


In some cases, saving final states without the complete model operating point can avoid some of the limitations of using the model operating point. For more information, see Save Block States and Simulation Operating Points.

Block Support

The following blocks do not support operating point:

  • In the Stack and Queue blocks, the default setting for the Push full stack option is Dynamic reallocation. This default setting does not support ModelOperatingPoint object. Other settings (Ignore, Warning and Error) support ModelOperatingPoint object.

  • Simscape™ Multibody™ First Generation blocks

Simulink tries to save the output of a block as part of an operating point. For S-functions, this happens even if the functions declare that no operating point is required. If the block output is of custom type, Simulink cannot save the operating point and displays an error. For more information about operating point use with S-functions, see S-Functions.

Model reference offers partial support for operating point. For details, see Model Referencing.


  • You can use only the normal or the accelerator simulation mode.

  • You cannot save operating point in normal mode and restore it in accelerator mode, or vice versa.

  • You cannot change the states of certain blocks that are not logged. For more information, see Change the States of a Block Within Operating Point.

Code Generation

The operating point feature does not support Simulink Coder™ or Embedded Coder® code generation.

Model Referencing

  • You cannot modify logged states of blocks that are inside a referenced model in accelerator mode.

  • Saving or restoring an operating point is not supported for model hierarchies that contain one or more referenced models configured to use a local solver. For more information, see Use Local Solvers in Referenced Models.

  • The following blocks do not support operating point when included in a model referenced in accelerator mode:

    • Level-2 MATLAB S-Function

    • MATLAB System

    • n-D Lookup Table

    • S-Function (with custom operating point or PWork vectors)

    • To File

Model Name Programmatic Interface

You cannot specify an operating point as the states input when you use the model name as a programmatic interface. For more information, see Use Model Name as Programmatic Interface.

See Also

Model Settings


Related Topics