Re: gEDA-user: LTSpice and gnucap

[al davis <ad136@xxxxxxxxxxxxxxxx>]

On Thursday 28 December 2006 07:27, Carlos Nieves wrote:
> > I'm trying to simulate a circuit with LTSpice, and I'm
> > getting an amazing simulation speed of 30ps of simulated
> > time by one real second, using a Pentium 4 machine at 3Ghz.

Why do you think that is amazing?

> > I want to try gnucap and see if it can go faster.

I would like to know too.  

My guess: maybe, maybe not.  There are lots of factors.  Just 
because one benchmark is faster on one than the other doesn't 
mean any other will be.  I suspect that large circuits will run 
faster on gnucap, small circuits will run faster on LTspice. 

I have one benchmark that runs in 40 seconds on gnucap, 8 hours 
on NGspice.  A couple of years ago I tried a similar one on 
LTspice.  I gave up waiting for it.  Gnucap ran in a few 
seconds.

Gnucap does well when it can optimize well, not so well when it 
can't.  Most Spice simulators don't try to optimize, but start 
from a faster base, by not having the overhead associated with 
optimization and mixed-mode.


> > I can generate the spice netlist, but as one of the
> > components I'm using is from Linear, its model is a .sub
> > binary file included with LTSpice.
> >
> > So I guess the .sub model is a propietary format and I
> > can't use gnucap for this simulation. Am I right?

Welcome to proprietary lockin.  That's what LTSpice is all 
about.



_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user