I developed a UVMF testbench previously. Now I want to use this environment as a subenv in a top-env. In top-env, transaction is sent to the predictor of the subenv(previous testbench). In addition, an unwanted transaction is sent from monitor_bfm of subenv to the predictor. I want to disable unwanted transaction from monitor_bfm of that agent to this predictor. How can I do this in the yaml files?
When re-using an environment, all the connections and agents/scoreboards/predictors should remain the same. The only thing that may change is an agent becoming passive as it may not be used to generate sequence items.
There shouldn’t be any other source of transactions to the sub-environment, so you will need to explain why you want to disable the functionality of the sub-environment.
Your new top environment may have a new scoreboard/predictor which takes sequence items from sub-environment components, but nothing should change in the sub-environment.
Thank you, yes, I could explain.
there is monitor_bfm in each agent which transform interface hdl signal to variable of transaction. It send transaction to predictor; It does not matter it was active or passive. When we use as subenv, the monitor in the agent still generate transaction and send it to the predictor. ( each variable has the value of zero). How can I prevent this systematically?
As I stated previously, you want your sub-environment to function exactly as it did originally. You want all of the monitors, scoreboards and checkers to be connected and function the same as when it is a standalone environment.
The only change may be that an agent may be passive instead of active, and this is controlled by the agent configuration generated by the test.
Why do you want to disable a monitor? This will break the functionality of the scoreboard/checker in the sub-environment.
I don’t need monitor and scoreboard. I only need to use the predictor in my testbench.
I want to remove from Tra in the uploaded diagram.
CamScanner 03-10-2025 20.59.pdf (278.1 KB)
If you don’t want any of the other components of the sub-environment, then you should just instantiate the predictor in your new environment.