“How do I model the clock signal?” – is a question frequently asked by hardware engineers who are new to using Simulink and HDL Coder. Here’s how it works:
- In Simulink, global signals such as clock, clock enable and reset are not explicitly modeled. Instead, they are created during code generation. You represent clock cycles in a Simulink model using sample time.
- For a single-rate model, 1 time step in Simulink maps to 1 clock cycle in HDL. You can use a relative mapping (e.g. a sample time of 1 = 1 HDL clock cycle) or an absolute mapping (e.g. a sample time of 10e-9 = one 10 ns clock cycle in HDL), depending on your preference and design requirement.
- For a multi-rate model, the fastest sample time maps to 1 clock cycle in HDL.
- Some optional optimization settings (e.g. sharing factor) and alternative block architecture (e.g. NewtonRaphson square root) introduce additional sample rates not present in the original model. In those cases, the fastest generated sample time is mapped to 1 HDL clock.
- Tip: Define sample time and ratio using MATLAB variables (e.g. Ts = 10e-9, upsamp = 4). This makes it easy to change all sample time settings quickly.
- If you model your sample time to be equal to the actual clock speed, that does not mean it will achieve this clock speed during RTL synthesis and implementation. Delays from logic gates and wires get introduced in synthesis and implementation, and the slowest path will dictate your fastest achievable clock frequency.
- HDL Coder can estimate some of these delays and provide direct feedback from RTL synthesis
