Re: [ng-spice] Inconsistent repeat noise analyses


To ng-spice@ieee.ing.uniroma1.it
From Reid van Melle <reid@analogsynthesis.com>
Date Mon, 8 Jan 2001 08:54:56 -0500
Delivered-To mailing list ng-spice@ieee.ing.uniroma1.it
In-Reply-To <Pine.BSI.4.05L.10101071558080.19283-400000@network.ucsd.edu >
Mailing-List contact ng-spice-help@ieee.ing.uniroma1.it; run by ezmlm
Organization Analog Design Automation
References <Pine.BSI.4.05L.10101071558080.19283-400000@network.ucsd.edu >
Reply-To ng-spice@ieee.ing.uniroma1.it

> In testing the problem I discovered a more glaring error, namely that the
> results of two identical noise analyses performed in succession do not
> necessarily yield the same results.  This sounds like an error caused by
> uninitialised variables or storage.  In fact I have just tried running the
> same analysis three times in succession, all noises in the second analysis
> are double those of the first, and in the third analysis are three times
> the noise of the first.  This pattern pattern is, however, not in itself
> consistent.

I'm not sure if my information is directly related to this problem or not.  
When I was working with the original Berkeley 3F5 source code (not sure how 
close the ng-spice noise analysis is to this), I found that the noise 
analysis was making use of uninitialized memory during its calculation.  I 
discovered this error when I was running the simulation engine in a loop and 
the noise analysis was causing segmentation faults.  I'm not sure why it does 
not segfault during a single run because the array indexing done in the noise 
analysis goes beyond the bounds of the malloc'ed memory.

Anyway, at the time I was more interested in getting the thing to run and not 
so much in getting correct results.  I simply extended some of the arrays and 
the problems disappeared.  I never had a chance to return to the problem to 
verify whether everything was correct.

                                        Reid

Partial thread listing: