Main Content

Get Started with TLM Generator

This example shows how to configure a Simulink® model to generate a SystemC™/TLM component using the tlmgenerator target for either Simulink Coder™ or Embedded Coder®.

For this example, we use a Simulink model of a FIR filter as the basis of the SystemC/TLM generation.

Requirements to run this example:

  • SystemC 2.3.1 (includes the TLM library)

  • For code verification, make and a compatible GNU-compiler, gcc, in your path on Linux®, or Visual Studio® compiler in your path on Windows®

Note: The example includes a code generation build procedure. Simulink does not permit you to build programs in the MATLAB® installation area. If necessary, change to a working directory that is not in the MATLAB installation area prior to starting any build.

1. Open Preconfigured Model

To open the FIR Filter model, click the Open Model button.

2. Set Simulink Coder Target to TLM Generator

a. Open the Configuration Parameters dialog box by selecting Simulation > Model Configuration Parameters in the model window.

b. In the Configuration Parameters dialog box, select the Code Generation view in the left-hand pane.

c. Under System target file, click Browse to select the TLM generator target. You can choose tlmgenerator_grt.tlc to use Simulink Coder or tlmgenerator_ert.tlc to use Embedded Coder for HDL Code generation. For this example, select tlmgenerator_grt.tlc.

3. Open TLM Generator View

In the Configuration Parameters dialog box, select the TLM Generator view in the left-hand pane.

The TLM Generator view has five tabs:

  • TLM Mapping

  • TLM Processing

  • TLM Timing

  • TLM Testbench

  • TLM Compilation

You will need to set different generator options in each pane.

4. Select TLM Mapping Options

In the TLM Mapping tab, Socket Mapping allows you to select the number of sockets for input data, output data, and control. Select the option, Three separate TLM sockets for input data, output data, and control.

The TLM Socket options allow you to define three different memory maps for your generated TLM component.

For this example, select the following for the input and output data sockets:

  • Auto-generated memory map for the input and output data sockets

  • Individual input and output address offsets for each data socket. This option generates a TLM component with one read register per model output and one write register per model input with individual addresses. Each Simulink model input is bound to its corresponding write register and each output is bound to its corresponding read register.

The other memory map options you may consider are:

  • No memory map: This option generates a TLM component with only one read and one write register without any address. The Simulink model inputs are bound to the write register and the outputs are bound to the read register.

  • [Auto-generated memory map with] Single input and output address offsets: This option generates a TLM component with only one read and one write register with one address each. The Simulink model inputs are bound to the write register and the outputs are bound to the read register.

When you generate the component with a memory map, you can add a command and status register, a test and set register, and tunable parameter registers.

For this example, select the following for the Control TLM Socket:

  • Include a command and status register in the memory map. The command and status register allows you to write command and read status during SystemC simulation; for example, manual buffering or buffer status.

  • Include tunable parameter registers in the memory map. Tunable parameter registers allow you to read or write the tunable parameter values.

Although not used in this example, a test and set register can be used as a mutex when multiple initiators access the component during SystemC simulation.

5. Select TLM Processing Options

In the TLM Generation pane, select the TLM Processing tab. The Algorithm Processing and the Interface processing options allow you to define different buffering and processing behaviors for your generated TLM component.

The algorithm execution options are:

  • SystemC Thread: The step function algorithm is executed in its own independent SystemC thread.

  • Callback Function: The step function algorithm is executed in a callback function called from the interface.

  • Periodic SystemC Thread: The step function algorithm is executed in its own time periodic independent SystemC thread.

The step function timing is determined by the value in the Algorithm step function timing (ns) field. The algorithm timing is counted with a wait() in the thread.

For this example, select SystemC Thread and enter 100 in the field for Algorithm step function timing.

Interface processing options:

  • Create an interrupt request port on the generated TLM component: This option creates an interrupt port (type signal, bool>) that is triggered every time a set of input has been processed.

For this example, select or enter the following choices:

  • Select Create an interrupt request port on the generated TLM component.

6. Select the TLM Timing Options

Select the TLM Timing tab. The Interface Timing section allows you to define the timing of the component input/output interface and processing thread.

For this example, the input and output delays are counted with wait() in the interface. Set a time value of 5ns for each of the transactions of Input Data Interface Timing, Output Data Interfacing Timing, and Control Interface Timing.

7. Select TLM Testbench View

Select the TLM Testbench tab. The TLM Generator target can generate a stand-alone SystemC/TLM test bench alongside the TLM component to verify the generated algorithm in the context of a TLM initiator/target pair. The TLM Testbench view provides run-time options for when the test bench code is generated and executed.

With the TLM Testbench options, you can:

  • Choose to see verbose messages echoed to the command window during the SystemC/TLM execution including TLM transaction and synchronization messages.

  • Indicate that the test bench should execute with or without timing annotations.

  • Indicate whether the initiator controls moving input and output datasets between the registers and the buffers or whether the component performs the moves automatically.

For this example, select Generate testbench, With timing for Run-time timing mode, and Automatic for both Input buffer triggering mode and Output buffer triggering mode.

After code generation has successfully occurred for a component and test bench, the Verify TLM Component button becomes enabled. Verify TLM Component performs the following:

  • Builds the generated code using make and generated makefiles.

  • Runs Simulink to capture input stimulus and expected results.

  • Converts the Simulink data to TLM vectors.

  • Runs the stand-alone SystemC/TLM testbench executable.

  • Converts the TLM results back to Simulink data.

  • Performs a data comparison.

  • Generates a Figure window for any signals that had data mis-compares.

The compilation of the generated files assumes the presence of make and a compatible GNU-compiler, gcc, in your path on Linux®, or Visual Studio® compiler in your path on Windows®.

8. Select TLM Compilation View

Select the TLM Compilation tab. This pane provides options to control the generation of makefiles used to compile the generated code.

Compiler Options:

The SystemC and TLM include library path options allow you to specify where the makefiles can find the SystemC and TLM installations. The default values allow you to use environment variables so that updates to your SystemC or TLM installations do not require updating your Simulink models. You can set up the environment prior to invoking MATLAB or use the MATLAB setenv command.

For this example, these are the environment variable values that were tested with standard OSCI installations located in /tools:

  • SYSTEMC_INC_PATH=/tools/systemc-2.3.0/include

  • SYSTEMC_LIB_PATH=/tools/systemc-2.3.0/lib-linux64

  • SYSTEMC_LIB_NAME=libsystemc.a (Linux) or systemc.lib (Windows)

  • TLM_INC_PATH=/tools/systemc-2.3.0/include

Toolchain:

On Windows this option allows you to select a compiler toolchain when multiple version of Microsoft Visual Studio are installed on the same machine. On Linux this option is fixed on gcc.

Component Naming:

This option allows you to add your own tag to the name of the generated component. The generated component name is built according to the following cases:

  • If a user tag is specified: modelname_usertag_tlm

  • If the user tag field is empty: modelname_tlm

For this example, enter intro for user tag.

9. Select Report

Select Report in the left-hand pane. For this example, select Create code generation report and Open report automatically. These options generate a html report during component generation. The Code generation report details the contents of each generated file.

10. Save TLM Generator Options

Click OK to apply these settings and exit the Configuration Parameters dialog box.

11. Build Model

In the model window, right-click on the DualFilter block and select C/C++ Code > Generate Code for this Subsystem in the context menu to start TLM component generation.

Alternatively, you can execute the following command in the MATLAB command window:

>> slbuild('tlmgdemo_intro/DualFilter');

During execution, you will be prompted to select the tunable parameters. The dropdown list of each coefficient allows you to select the storage class of the variable. The Storage Class options are:

  • Inlined - The inlined parameters are not tunable.

  • SimulinkGlobal - The SimulinkGlobal variables are tunable.

ExportedGlobal, ImportedExtern and ImportedExternPointer are not supported by the TLM Generation model.

The option Simulink Global has been selected for this example. Click Build. The TLM generation is completed when you see the following message appear in the MATLAB command window:

### Starting Simulink Coder build procedure for model: DualFilter
### Successful completion of Simulink Coder build procedure for model: DualFilter

Build Summary

Top model targets built:

Model       Action           Rebuild Reason                                    
===============================================================================
DualFilter  Code generated.  Code generation information file does not exist.  

1 of 1 models built (0 models already up to date)
Build duration: 0h 0m 16.579s

12. Open Generated Files

Open the generated files in the MATLAB web browser by clicking on the links in the generated report or in the MATLAB Editor (the generated files and report are located in your current working directory):

13. Verify Generated Code

a. Open the Model Configuration Parameters dialog box by selecting Simulation > Configuration Parameters in the model window.

b. In the Configuration Parameters dialog box, select the TLM Generator view, and then select the TLM Testbench tab.

c. In the TLM Testbench pane, click Verify TLM Component, to run the generated testbench.

Alternatively,you can execute the following command in the MATLAB command window:

  verifyTlmgDemoModel('intro')

This verify step performs the following actions:

  • Builds the generated code.

  • Runs Simulink to capture input stimulus and expected results.

  • Converts the Simulink data to TLM vectors.

  • Runs the stand-alone SystemC/TLM test bench executable.

  • Converts the TLM results back to Simulink data.

  • Performs a data comparison.

  • Generates a Figure window for any signals that had data mis-compares.

14. Review Execution Log

The option to generate test bench allows you to see how the test bench initiator threads interact and synchronize with the target. Look for the comparison result at the end of the log and verify that the data comparison is successful.

### Comparing expected vs. actual results.
Data successfully compared for signal tlmg_out1.
Data successfully compared for signal tlmg_out2.
### Component verification completed

This concludes the Getting Started with TLM Generator example.