Re: [ng-spice-devel] Re: snapshot/resimulate.


To ng-spice-devel@ieee.ing.uniroma1.it
From Al Davis <aldavis@ieee.org>
Date Wed, 11 Apr 2001 22:50:04 -0700
Delivered-To mailing list ng-spice-devel@ieee.ing.uniroma1.it
In-Reply-To <200104041534.IAA91472@spnet.com >
Mailing-List contact ng-spice-devel-help@ieee.ing.uniroma1.it; run by ezmlm
References <200104041534.IAA91472@spnet.com >
Reply-To ng-spice-devel@ieee.ing.uniroma1.it

On Wed, 04 Apr 2001, Ed Hudson wrote:
> .....large circuits may have many nodes with little
> activity - a ram, for example, might have 100 active nodes and a
> million stable nodes...

Spice does nothing to take advantage of this, except for a very 
primitive bypass in the transistor code.  ACS seriously takes 
advantage of this, but not as much as future versions will.

Stable parts of the circuit are not processed every iteration, since 
this is driven by queues.  They are not even loaded into the matrix, 
and if whole blocks are stable, the stable parts of the matrix are 
not solved.  It just keeps the old values.  There are two notions of 
stability here ...  A circuit might be locally linear with moving 
voltages.  In this case, it can still take the matrix shortcut, but 
not the full evaluation shortcut.

The "multi-rate" in the other mail I sent is that some sections might 
be moving slowly compared to others.  Moving slowly is still moving, 
so ACS will simulate it, for now.


Partial thread listing: