Regarding the Sequences/Slave page in the Cookbook, Verilab published a paper at SNUG 2015 with a superior (IMO) approach, where the monitor decodes requests (as typical in UVM agents) and the sequencer then sends only responses to the driver.
In reply to sjohnson-aril:
I read both approaches.
For the method mentioned in this paper, can someone share the source code, please?
I honestly didn’t understand why the p_sequencer appears in the code snippet of the slave response sequence.
Also, I won’t call it a superior approach.