The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Jul 20th, 2024, 4:24am
Pages: 1
Send Topic Print
Different results for same netlist (Read 3662 times)
Cri Azzolini
Community Member
***
Offline



Posts: 48
University of Parma, Italy
Different results for same netlist
Jan 26th, 2006, 1:44am
 
Hi all,

I would like to submit a question regarding Spectre simulator.
Sometimes it happens to me that an analog design netlisted as a flat circuit (flat netlist) produces a certain output whereas the SAME circuit netlisted using hierarchical schematics produces different results. I heard this queer behaviour happened also to colleagues of mine but I would like to understand the reason.
Can anyone explain this point?

Thanks in advance,
Cri
Back to top
 
 
View Profile WWW cri.azzolini   IP Logged
Paul
Community Fellow
*****
Offline



Posts: 351
Switzerland
Re: Different results for same netlist
Reply #1 - Jan 27th, 2006, 2:21pm
 
Hi Cri,

as I am using hierarchical netlists by default, I haven't experienced such problems. Intuitively, I would tend to blame the netlister and not the simulator. In any case, make sure the netlister renetlists all the blocks, these guys having a bad habit to keep older parts somewhere in memory to reduce the netlisting duration.

Paul
Back to top
 
 
View Profile WWW   IP Logged
Cri Azzolini
Community Member
***
Offline



Posts: 48
University of Parma, Italy
Re: Different results for same netlist
Reply #2 - Jan 28th, 2006, 12:42pm
 
Hi Paul,

I am going to follow your advice and I will check for a complete re-netlisting. I am also quite sure that the netlister is the cause but this means that the simulator (in this case Spectre) can solve the circuit's equations depending on how the circuit is described (i.e. flat vs hierarchical)! It seems to me very strange since I can not imagine how this is possible: I mean, the equations are the same and the initial condition are (I suppose) the same.

In the meanwhile I will try to find an idea but I would be grateful if anybody out there had an explanation.

Bye,
Cri
Back to top
 
 
View Profile WWW cri.azzolini   IP Logged
Paul
Community Fellow
*****
Offline



Posts: 351
Switzerland
Re: Different results for same netlist
Reply #3 - Jan 29th, 2006, 1:12pm
 
Hi Cri,

what I was trying to say is that IMHO the netlister doesn't generate a correct flat netlist and that's why the simulation results differ. I don't think Spectre does any difference on this...

Paul
Back to top
 
 
View Profile WWW   IP Logged
Andrew Beckett
Senior Fellow
******
Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK
Re: Different results for same netlist
Reply #4 - Jan 30th, 2006, 2:09am
 
There is always the possibility that you have a circuit that has multiple operating points - and that the difference in order of processing is causing the numerical algorithms to prefer one solution over the other (i.e. both are correct, but there's an element of luck to which solution you'll get).

However, the most likely explanation is that the circuit is not in fact identical. I don't know how you're generating a flat netlist (this used to be supported in the spectreS interface, but not in the newer spectre interface) - it might just be that you have two different schematics - in which case perhaps you entered something wrongly.

I'd compare the netlists and see if you can spot any difference.

Best Regards,

Andrew.
Back to top
 
 
View Profile WWW   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.