[SOLVED] VCS "X"state problem

Hi all,
When i use vcs to simulate a design with behavioral models,some "x" or "z" state found especially some read only registers.How should I deal with them more effectively?Because there are too many....anyone can guide me ?
thanks

rettylee said:

Hi all,
When i use vcs to simulate a design with behavioral models,some "x" or "z" state found especially some read only registers.How should I deal with them more effectively?Because there are too many....anyone can guide me ?
thanks

Click to expand...

1): For "Z" value, it must be defined but not assign some value. You can choose to clear your RTL code or just let it there.
2): For "X" value, it maybe:
a): multi-driven case. //you need fix it in RTL code
b): regsiter without reset. //you safely let it there if the function is right.

yx.yang said:

1): For "Z" value, it must be defined but not assign some value. You can choose to clear your RTL code or just let it there.
2): For "X" value, it maybe:
a): multi-driven case. //you need fix it in RTL code
b): regsiter without reset. //you safely let it there if the function is right.

Click to expand...

Thank you for your reply,yang.
as you say, I think most "x" came from registers without reset or behavioral models without initial value.
To fix them, now,my only choice that I can chose is to force the signal one by one.I want to kown ,is there better method to deal with it?
guide me please!

rettylee said:

Thank you for your reply,yang.
as you say, I think most "x" came from registers without reset or behavioral models without initial value.
To fix them, now,my only choice that I can chose is to force the signal one by one.I want to kown ,is there better method to deal with it?
guide me please!

Click to expand...

Hi, rettylee:
I think you have some confuse on the purpose of simulation now.
Just think about why you need do simulation? Is it for just let you see the simulation result is pass or to verify the RTL code(or your behavior model) is function right or wrong??? Just think about it first.
So, at RTL(or your behavior model) simulation stage, I don't think there is any need to force any signal to some value. Just let the simulation behavior as it should be. If there is function fail, you need fix it in your RTL code or behavior model, not just force it in simulation.
Thanks.

Thank you for your guide,it's clear for me now.