Difference between Master and Slave Agents

[Shankhadeep]

Can anyone please explain what is the difference between Master agent and slave agent? I have tried searching through many blogs and guides, but could not come to a final conclusion.

In my Testbench architecture, I am verifying a SECDED Hamming code Encoder-decoder. These are the steps

1. Agent 0 drives Data_in and valid to Encoder.

2. Agent 1 monitor receives encoded data with parity and passes it to Agent 2 (re-active agent)

3. Agent 2 generates error injection and drives the error-injected data with parity to decoder.

4. Agent 3 captures the type of error (single/double) from Decoder.

5. Agent 4 monitor captures the decoded data and passes it to scoreboard.

 

I am confused here, which agents are masters and which are slaves. 

 

[apfitch]

Hi, in UVM people tend to talk about agents as being passive, active, or reactive instead of "master" or "slave".

Each agent contains a driver, a monitor, and a sequencer.

An active agent has all three.

A passive agent monitors, but does not drive.

A reactive agent drives, but depends on the DUT more closely (i.e. the agent drives data to the DUT, but reacts in some way with the DUT.

So I guess you could be using "master" to mean active, and "slave" to mean reactive *or* passive.

regards

Alan