Re: [ng-spice-devel] [ng-spice] ACS, model compiler, cross-post from gEDA list (fwd)


To Al Davis <aldavis@ieee.org>
From Steve Hamm <Steve.Hamm@motorola.com>
Date Fri, 8 Sep 2000 16:24:14 -0500 (CDT)
Cc ng-spice-devel@ieee.ing.uniroma1.it
Delivered-To mailing list ng-spice-devel@ieee.ing.uniroma1.it
In-Reply-To <00090811505602.06449@oliver.al.dynip.com >
Mailing-List contact ng-spice-devel-help@ieee.ing.uniroma1.it; run by ezmlm
References <Pine.LNX.3.96.1000908075931.8820D-100000@ieee.ing.uniroma1.it ><14776.61799.312611.578899@pearls ><00090811505602.06449@oliver.al.dynip.com >
Reply-To ng-spice-devel@ieee.ing.uniroma1.it

---"Al" == Al Davis <aldavis@ieee.org> writes:

Al> Mine doesn't do automatic differentiation.

That's unfortunate. Errors in derivatives are the one of the main
reasons for nonconvergence in SPICE, right in there with
discontinuities in the i-v and q-v characteristics. Automatic
differentiation at least gives a correct (almost always) set of
derivatives to start with.

Al> Fine, but it still doesn't handle the nuisance stuff, and translating
Al> existing models to Verilog-A probably won't happen.  Mine is designed
Al> so that transating the existing models to it is equivalent to
Al> translating to a slightly different architecture.  Then it opens up
Al> the ability for bigger architecture changes.

Translating most existing models to Verilog-A is fairly trivial; even
so, a production model will want some hand-tweaking to eliminate that
>3x penalty for automatic differentiation. But you're right, the
boilerplate for handling all the standard add-ons can get tedious. I'd
be curious to see how you've worked this out.

--Steve


Partial thread listing: