Validate HDL Implementation Model to Simscape Algorithm

If you design your algorithm by using Simscape™ switched linear blocks, you can run the Simscape HDL Workflow Advisor to generate an HDL implementation model. The HDL implementation model represents the Simscape algorithm by using Simulink® blocks that are compatible for HDL code generation.

Before you prototype the implementation model on an FPGA or target Speedgoat FPGA I/O modules, you can verify the functionality of your design in the Simulink modeling environment. To verify the functionality, specify insertion of validation logic in the HDL implementation model when you run the Simscape HDL Workflow Advisor. This logic verifies whether the numeric results of the HDL implementation model match the original Simscape algorithm.

In some cases, there can be a mismatch in simulation results between the Simscape algorithm and the corresponding HDL implementation. Such mismatches generate warnings or assertions when you simulate the implementation model. To resolve the warnings, use a combination of various settings in the Generate implementation model task as illustrated below.

Bridge Rectifier Model

This example uses the bridge rectifier model to illustrate how to generate an implementation model with validation logic inserted in the model, and how you can resolve any assertions that may be generated when you simulate the implementation model.

  1. Open the bridge rectifier model. In the MATLAB® Command Window, enter:


    Inside the Simscape_system, you see four diodes arranged in a bridge configuration. For both positive and negative input values, this configuration provides a positive, rectified output.

  2. Open the Simscape HDL Workflow Advisor for your model:

  3. Right-click the Get state-space parameters task and select Run to Selected Task to run all tasks in the Advisor except for the Generate implementation model task.

  4. In the Generate implementation model task, select the Generate validation logic for the implementation model check box. Leave the other options with their default values and select Run This Task.

    After running this task, keep the UI window for this task open. If simulating the HDL implementation model generates warnings, you modify the settings in the Generate implementation model task and then rerun this task. You do not have to modify or rerun other tasks.

  5. Click the link to open the HDL implementation model. You see a Validation Subsystem that compares the simulation results of the Simscape model to the HDL implementation model. Simulate the implementation model.

You see that simulating the model generates multiple assertions indicating a mismatch in the simulation results. If you open the Diagnostic Viewer, you see this message:

Assertion detected in 'gmStateSpaceHDL_BridgeRectifier_HDL_SimMismatch/Validation/Check Static Range1' at time 0.04186 [4982 similar]

The message indicates that the Simscape™ algorithm does not match the equivalent HDL implementation. To resolve the validation mismatch, you can modify various settings in the Generate implementation model task until the HDL implementation model matches the Simscape algorithm. In most cases, to resolve the numeric mismatch, you may want to use a combination of these settings.

Increase Validation Logic Tolerance

Conversion of a Simscape algorithm to an equivalent HDL implementation leads to rounding errors. The default tolerance value is relatively small and can be difficult to achieve especially with single-precision data types in the HDL implementation model. To resolve the mismatch:

  1. Start by increasing the Validation logic tolerance to an initial value such as 1e-4.

  2. Select Generate validation logic for the implementation model and run the task to generate the HDL implementation model that includes validation logic.

  3. Simulate the model and check whether the simulation displays assertions in the Diagnostic Viewer. If the simulation results produce warnings, proceed to the next step to increase the number of solver iterations.

Increase Number of Solver Iterations

For each mode in the physical system, the switched linear workflow arrives at a state-space representation. The solver method is iterative and performs multiple computations to determine the correct mode for the next time step. After a certain number of iterations, the output value from the next time step becomes the same as the value from the previous time step. This consistency in the output value indicates the correct number of solver iterations.

The Advisor by default chooses an optimal value for the number of solver iterations. See Using Number of Solver Iterations. If increasing the tolerance value does not improve accuracy of the HDL implementation model, you can resolve the numeric mismatch by increasing the number of solver iterations.

When you increase the number of solver iterations, the code generator changes the sample time of the generated HDL implementation model. A large number of iterations can increase the simulation time significantly. See Reducing Number of Solver Iterations. This flowchart illustrates how to change the Number of solver iterations.

Use Larger Floating-Point Precision

You can use the Floating-point precision setting in the Generate implementation model task to specify the floating-point data type you want to use for the algorithm inside the HDL Subsystem. Specify whether you want to store the matrix coefficients in single or double data types and whether to use single or double when performing the computations.

Floating-Point PrecisionDescription
DoubleUsing double floating-point precision increases the numerical accuracy of the generated model and the maximum achievable target frequency. However, the area consumption and pipeline latency are also increased.
SingleThis is the default setting for floating-point precision.
Single coefficient, double computationThis mode offers a tradeoff between Single and Double modes of floating-point precision. To save memory usage, the coefficients that are stored in single. The matrix computations are then performed in double for improved accuracy.

This flowchart illustrates how to change the Floating Point Precision and improve the numeric accuracy of the generated HDL implementation model.


Double-precision operations have large latencies and require a large Oversampling factor to allocate sufficient delays for the floating-point operations, which reduces the sampling frequency. For a tradeoff between accuracy and precision, use Single coefficient, double computation as the Floating Point Precision.

After specifying double data types, if the simulation results still produce warnings:

  1. Proceed to the first step to further increase the validation logic tolerance. Use a tolerance value of 1e-03 and then simulate the model to see if the numeric accuracy requirements are met.

  2. Increase the number of solver iterations if you still see warnings in the Diagnostic Viewer. Continue iterating between these steps till the HDL implementation model numerically matches the Simscape algorithm.

For the bridge rectifier model, to resolve the warnings, set the Validation logic tolerance to 1e-4 and specify the Floating Point Precision as double. After you generate the implementation model with the validation logic, you see that simulating the model does not display warnings in the Diagnostic Viewer.

See Also


Related Topics