Compiling Berkeley SPICE 3f4
Saravanan Solaimalai
machine using DevC++'s gcc compiler. I've looked into the batch files
for compilation with Microsoft C 5.1 and tried translating it so that I
can use DevC++'s porting of the GNU tools but I got stuck with several
things. Has anyone attempted to compile this Spice code with anything
similar, i.e. GNU tools ported for Windows?
HelpMe
Mike Engelhardt
of date with the project. The makefiles are a better clue.
But you're probably better off if you remake the project
yourself, at least that's been my experience.
Back in '91 or there abouts, I was the first person to port
Berkeley SPICE to Linux, which used gcc, so you might be
able to find my spice3e2.kit.TZ file somewhere which will
have lot's of little problems fixed.
But for a PC based SPICE, there's nothing better in terms
of speed and convergence than the SPICE from Linear
Technology called SwitcherCAD III availible for free
from www.linear-tech.com/software. It includes an
integrated schematic capture and marching waveforms, too.
Good luck,
Mike Engelhardt
"Saravanan Solaimalai" <Saravanan....@umist.ac.uk> wrote in message
news:3B77CECF...@hotmail.com...
Kevin Aylward
"Saravanan Solaimalai" <Saravanan....@umist.ac.uk> wrote in message
news:3B77CECF...@hotmail.com...
I can give you the full source code of my Windows M.S. Visual C++ port of
the XSpice engine that I use in my SuperSpice if you want. You just load in
the project file and your done. email me.
Kevin Aylward , Warden of the Kings Ale
ke...@anasoft.co.uk
http://www.anasoft.co.uk - SuperSpice "Cheap, No Shit!",
GUI xspice, an affordable unlimited component, mixed-mode Windows simulator
with Schematic Capture, waveform display, FFT's and Filter Design.
Opinions of my employer are not necessarily indicative of my own
Oscillators don't, amplifiers do"
Ken Ingram
Pedro F Giffuni
Mike Engelhardt wrote:
>
> I wouldn't use the batch files. I'm pretty sure they're out
> of date with the project. The makefiles are a better clue.
> But you're probably better off if you remake the project
> yourself, at least that's been my experience.
>
> Back in '91 or there abouts, I was the first person to port
> Berkeley SPICE to Linux, which used gcc, so you might be
> able to find my spice3e2.kit.TZ file somewhere which will
> have lot's of little problems fixed.
>
Well... your port was re-ported to FreeBSD, and when the original Spice
was made available from Berkeley, most of your patches were moved to the
FreeBSD distribution. There is a set of patches from other people that
are not yet integrated as they became too big to maintain apart from
Berkeley's spice. My WIP to update this is here:
http://www.pitt.edu/~pfg1/spice/
I haven't properly added all the devices due to lack of time, but this
should build on any UNIX. I was tempted to start a port to MS Visual
C++ but I don't have the time anymore. Still, I'm sure there are memory
leaks and other nasty bugs, so I would recommend OPUS Spice which,
although doesn't carry sources, integrated UGA's XSpice and made many
fixes.
cheers,
Pedro.
Mike Engelhardt
I've never used OPUS, but you might be missing a bet if you
don't try the Linear Tech SPICE I wrote. Though it also
doesn't come with source code, it is a complete reorganization
of the Berkeley SPICE to improve it's performance on machines
with faster processors than memory. It does P-SPICE
semiconductive models, like bipolar quasi-saturation,
recombination current, and deep impact ionization. Also,
it can do legacy behavioral modeling(using POLY()) as well
as general expressions, look-up tables and Laplace. The
waveform files are compressed and demand paged into memory.
A typical simulation might generate 8GB of raw data compressed
on the fly to a 400MB disk file -- but you'd only use 60MB
to view it. It does BSIM3.2.2, BSIM4.1 and a MOSFET device I
wrote to handle the unique charge behavior of the discrete power
VDMOS. It has integrated schematic capture, but can open and
edit SPICE netlists(size limited by virtual memory). IC
design engineers use it as an SPICE accelerator for full-chip
transistor level simulation. All without leaking one byte
of RAM:)
It doesn't do xspice devices, but that's because I've had too
much trouble with timestep control in xspice, so I wrote a
special mixed mode simulator rooted in predictor/corrector
methods which runs circles around xspice.
Again, it's called SwitcherCAD III from
www.linear-tech.com/software/ It's business model is sales
collateral, so it's free. But to succeed in this model for
selling switch mode power supplies, it pretty had to be the
worlds fastest SPICE.
Mike
"Pedro F Giffuni" <pf...@pitt.edu> wrote in message
news:3B7AD9F3...@pitt.edu...