Unknown variable appear in ADE window when I run the simulation

Dear Mostafa,

Mostafa A said:

When I run the simulation, some unkown variable appear in ADE window (circled in the attached pic). Parameters are "intf", "pf", "pPar", "pP", "p", "pw", "e", "nf", "n", "nw". Do you know what these variable are and how should I get rid of them?

As you probably know, Cadence no longer supports ADE-L. I think it would be worth your while to migrate to ADE Assembler/Explorer. Explorer is the equivalent of ADE-L and its feature set is far richer than ADE-L. The migration is quite easy as the two GUI are similar.

It appears these variables may be associated with parameters of your devices just from their names - but this is just a guess. Do their presence preclude you performing a transient simulation in IC 6.1.5? If not, I would ignore the design variables and proceed with your simulations in IC6.1.5, but perhaps check if the simulation results you obtain in 6.1.7 are identical to those in IC 6.1.5.

If the presence of the unassigned variables preclude you from running in IC6.1.5, I would debug the issue using the following methodology:

1. In IC 6.1.5, create a netlist (Simulation->Netlist-Recreate (See Figure 1)

2. Navigate to the netlist directory and make a copy of the input .scs file as, for example, 6p1p5_input.scs

3. Close your session and open a new session in IC 6.1.7

4. In IC 6.1.7, create a netlist (Simulation->Netlist-Recreate

5. Navigate to the netlist directory and make a copy of the input .scs file as, for example, 6p1p7_input.scs

6. Compare the two files 6p1p5_input.scs and 6p1p7_input.scs (UNIX diff, or visual editor tkdiff)

   a. Examine each and search either for the "unknown" design variable names shown in your screen capture ADE-L window.

   b. Are the design variables just in one of the two netlists? 

   c. Are the design variables assigned a value in either of the two netlists?

7. Understanding the "unknown" design variables in the two netlists will provide you information about whether their values will impact simulation results. I do not know of a command to "ignore" selected design variables. I suspect one does not exist as if a design variable that is used in a simulation will prevent spectre from computing some element in the netlist...and hence the simulation can not be performed.

I hope this helps Mostafa. Others in the forum may have better or more more information on your problem.

Shawn

Dear Shawn, thank you for reply.. I forgot to mention, the simulation "terminate prematurely due to fatal error"... and you referred to "Figure 1" in you comment.. I think the figure is not attached.

Dear Mostafa,

Mostafa A said:

I forgot to mention, the simulation "terminate prematurely due to fatal error"..

If that is the case, I might suggest the process I proposed if you think it worth it to better understand the "unknown" design variables. 

Mostafa A said:

.. and you referred to "Figure 1" in you comment.. I think the figure is not attached.

You are right! Sorry, and  It is now attached.

Shawn

Figure 1

Thank you dear Shawn

My guess is that a pPar() expression ended up in the netlist (which it shouldn't do normally) and then spectre failed because it couldn't find "variables" used in this erroneous expression, and then ADE added them to the design variable list. Knowing how spectre failed in the first place would be useful (delete the variables in ADE and run again, and then show the error and ideally the line in the netlist that the error is referring to).

Is $CDS_Netlisting_Mode set? In the CIW, try typing:

cdsGetNetlistMode()

If this returns "Digital" (not "Analog") then I suggest in the terminal you started virtuoso from, type;

setenv CDS_Netlisting_Mode Analog

or

export CDS_Netlisting_Mode=Analog

(first for csh, second for bash/ksh)

then start virtuoso again in that terminal window (exit any you are running) and see if that cures it (you'll need to delete any of the variables that ADE has added first). There's another thing that could cause this, but that would mean something was wrong with the CDF which would seem unlikely if it's the PDK, and also would affect both IC615 and IC617.

BTW, Shawn's advice about moving to Explorer/Assembler is good, but you can't do that if you're using IC615 anyway.

Regards,

Andrew

Thank you dear Andrew,

cdsGetNetlistMode() returns "Analog".

The problem is solved actually. The unknown parameters appeared in the ADE were related to the prepared inverter in 65nm TSMC library (bottom one in the attached photo).. While such inverters have their default parameters values, the parameters will appear in ADE window!!!!     Anyway, I replaced it with the inverter from the 65nm standard cell library (the top one in the photo) and now I don't have problem..Thanks.