Hi verification experts,
This is a question about methodology.
I have a RW register (called Areg) that is updated by writing other register (called Breg) of the DUT and also by writing directly this registers.
I have created a quirky register description for Breg and each time “Breg register” is written it makes a prediction/update of Areg as expected.
I have seen that each time Breg register is predicted then it destroys the randomization created/done in my stimuli sequence for Areg. In my opinion, that a register prediction has an influence in the testbench stimuli sequence is wrong. (from the methodology point of view)
Is there any reason UVM doesn´t provide a function that only changes/predicts the “m_mirrored” value and leave intact the .value variable?
is there any reason to force the predict function to update “m_mirrored” value, “m_desired” value and randomize “value” variable?
My view is that a scoreboard/register prediction should not influence anything in the stimuli sequence generation, but that is not the case.
I am looking forward your answers.