verilog VPI回电

Question

What is VPI call back cbValueChange supposed to do and how does it work?

Suppose I want to execute a function at the posedge of signal X (about 2000 cycles). Should I use call back cbValueChange or call the code from the test bench?

always @posedge(X) 
begin 
    $pli 
end 

Which would be more efficient?

Answer 1

cbValueChange is the reason used to register/receive value change event for a registered signal via vpi_register_cb() . So when the signal value got flipped or changed, the simulator will invoke the callback function at the proper simulation phase. If you want to invoke your own PLI/VPI function only, you can see the detail example in http://www.asic-world.com/verilog/pli5.html

Answer 2

Efficiency wise it probably doesn't matter much. Without a better understanding of your situation it's tough to give solid advice but I recommend the $pli call as in your example. Using cbValueChange means registering some C code to be executed by the simulator when the value of a variable changes. This happens without a user defined $pli routine being called at the time the value changes. Since this is completely invisible to anybody reading the Verilog code, you will frustrate anybody trying to understand what is going on if your C code makes any changes that affect the simulation.