RE: PSPICE format


To "'ng-spice-devel@ieee.ing.uniroma1.it'" <ng-spice-devel@ieee.ing.uniroma1.it>
From "Gillespie, Alan" <Alan.Gillespie@analog.com>
Date Thu, 7 Dec 2000 10:18:24 -0000
Delivered-To mailing list ng-spice-devel@ieee.ing.uniroma1.it
Mailing-List contact ng-spice-devel-help@ieee.ing.uniroma1.it; run by ezmlm
Reply-To ng-spice-devel@ieee.ing.uniroma1.it


> I think I will 
> reverto the output
> to the standard spice3 raw format in rewok14 and make the 
> probe compatible
> format as an option (a file to substitute). 

I've a feeling that we should just remove it. When I submitted
my "fixes" to Paolo, most of them were surrounded by
#ifdef AlansFixes statements, so it was easy to pick and choose
either the new code or the old code to include in ng-spice. 
However, the changes I'd made to outitf.c were some of the first
I ever did, and were done before I got into the habit of leaving
the old code in a working state.

I think Paolo reckoned that some of the changes in outitf.c were
worth keeping, and since the only quick way to do it was to
incorporate all my changes, that's what happened. (Correct me if
I'm wrong Paolo).

It's really my lack of effort in helping Paolo and Arno with these
changes which means that we haven't removed the PSPICE stuff
already. (I won't bore you with excuses).

The reason I originally put the PSPICE data format in spice, was
that I was running on a PC, and I didn't like the spice viewing
tool. I'd been used to using the PSPICE tools for years at work,
and was quite comfortable with them, and I also knew that Microsim's
demo version of PSPICE included a version of Probe which was not
limited by filesize (the simulator was, but not the viewer). I could
therefore run Spice3f4 in batch mode, and use the demo version of
Probe to look at the results (so long as I put a little code in the
data file to convince Probe that it had been produced by the demo
version of PSPICE).

I was originally given the documentation for the Probe data file
format by Microsim, for "legitimate" reasons, i.e. we were writing
tools to post process probe files - reducing file sizes, discarding
data, searching for stuff, etc. I'm very grateful to Microsim (as
they were then) for giving out this information, and we shouldn't
be using it "against" them. So I don't really want to pass on the
documentation without checking with "Mr. PSPICE" first. Since Microsim
were bought by Orcad, who were bought by Cadence, we're now talking
about a whole different company, who don't have such a good
reputation for being "nice" to deal with. Maybe it's worth getting
in touch with them, though.

There are more practical, and less "legal" reasons to remove my
outitf.c changes, though. First of all, since ng-spice doesn't run
on the PC (I mean win32), I don't know how much use the PSPICE format
would be to ng-spice. Next, the other main fix that I can think of
in outitf.c is the one which removes all the device internal nodes,
and optionally replaces them with terminal currents. They way I did
this is a horrible hack, (as I said, it was early days), and I think
it could be re-done in a much better manner. Lastly, the PSPICE format
is just another format, it's not much different from most other formats,
although it probably does combine the best bits out of most of them.
Anyway, I think it could be vastly improved, so I'll send another email
explaining what I think we could do to create a new ng-spice format
which beats most of the others.

So what do people think ? We can't keep it in without getting approval,
but is it worth a great deal of effort to get that approval ?

Cheers,

Alan

Partial thread listing: