[SI-LIST] Re: Historical question: HSpice from Avant! ?

  • From: "Chan, Michael" <Michael.Chan@xxxxxxxxxx>
  • To: <chris.cheng@xxxxxxxxxxxx>
  • Date: Wed, 31 Oct 2001 13:31:30 -0600

Chris:
      IBM one is called "ASTAP".

Michael

-----Original Message-----
From: Chris Cheng [mailto:chris.cheng@xxxxxxxxxxxx]
Sent: Wednesday, October 31, 2001 1:23 PM
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Historical question: HSpice from Avant! ?



Jim,
Good piece of history. For the longest time I thought Aspec was
IBM proprietary and it runs on statistical model rather than
single point best/worst case parameters.
Chris

-----Original Message-----
From: jim freeman [mailto:kacief@xxxxxxxxxxxxx]
Sent: Wednesday, October 31, 2001 9:52 AM
To: Andrew.Ingraham@xxxxxxxxxx
Cc: jkolstad@xxxxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Historical question: HSpice from Avant! ?



Hi All,
    Before Hspice, there was no industry standard.All of the IC vendors
had
their own versions that were derive from Berkeley spice or home grown.
Another version of Spice was Aspec(some people mistakenly call it
Aspect).
It was written by independents that rented it out thru the industry at
very
high prices. It was a MIPS hog beause it worked on conservation of
charge.
There was a competing version that was called Ispice that used the basic
algorithms employed by Berkeley Spice. The Spice versions were not
charge
conservative. That is why there are outrageous voltages sometimes
generated
by a spice simulation. Try to sim an oscillator wih a 5V power supply
and
watch the voltages soar.
    Somewhere around 1980-84, the writers of Aspec, which was full
featured
and very expensive to use( the source code was protected and simulations
could only be done over a teletype terminal) both because of dialup fees
and
the MIPS required to run, sold the source code to Intel and quit issuing
upgrades.
    Enter Sean and Kim Haley of metasoft with Hspice. They decided to
generate all of the features of Aspec( they were prompted to do so
because
the users wouldn't use their simulator withou the features). One of the
early options that was adopted was .option wl. This would allow the use
of
the  n and l parameters in reverse order. This was an option that was
povided by Aspec and was very useful because some IC vendors provided
the
values in reverse order than other IC vendors. A designer was forced to
perform the reversal by hand and then simulate(imagine the typing errors
that that caused and the wrong simulations). At thistime the l=3D and =
W=3D
were
not available in the code. Sean and Kim also drastically reduced the
cost
over the Aspec simulator and also the run time for simulations. They
also
allowed the source ode to be sold to some vendors in the mid 80's so
that
the vendors could ad features. Most of these features were then aded to
the
updated source code and a standard was born.
    The system manufacturers started demanding Spice like simulations in
the
late 80's and the standard that was picked as a defacto was Hspice
because
of its rich feature set and openness of the writers to add more
features.
    One of thefeatures that made Hspice especially useful in board and
system simulations was the addition of the dtemp parameter. If it is
added
to every transistor in the netlist, the temperature of some transitors
can
be different than others in the simulation.

Thanks
Jim Freeman

"Ingraham, Andrew" wrote:

> > wondering... how did HSpice by Avant! seem to become the industry
> > standard?
> > (Many simulators advertise themselves as "HSpice compatible.")
Since
> > Avant!
> > has only been around since 1995, where did HSpice come from?
> >=3D20
> HSPICE came from Meta Software (I think that was the one ... it was
one
> of a few companies called Meta-something) before Avant! acquired it.
> HSPICE had already been around for years.
>
> I think that one of the things that made it a defacto industry
standard,
> is that they seem to have bent over backwards to add feature after
> feature.
>
> Among those features, were dozens of proprietary transistor models
> (MOSFET LEVEL=3D3D*), which exist nowhere else but HSPICE.  Some of =
them
=3D
> are
> unique to a company's FAB line, many have since been dropped.  Others,
> like the BSIM3 level=3D3D47, have themselves become defacto standards;
not
> just because it's BSIM3, but because the level=3D3D47 implementation =
=3D
> "caught
> on."  If I am an IC vendor and I use HSPICE's level=3D3D47 models, and
you
> are one of my many customers and you want to simulate around my IC,
you
> are forced to use HSPICE for those simulations.
>
> Unfortunately, HSPICE's feature-laden characteristics also make it
> difficult to learn and understand (compare how many .OPTIONS choices
> HSPICE has vs. traditional Berkeley SPICE).  Furthermore, Meta and
> Avant! have a habit of tweaking the algorithms every few releases
which
> can cause plenty of problems.  But since you're already locked in to
> using HSPICE (on account of its proprietary models), you learn to put
up
> with them, to be on the lookout for subtle (and not-so-subtle!)
changes,
> and to work around them.
>
> Some of the descriptions in my HSPICE manual are flat-out wrong,
because
> Meta/Avant! changed the algorithm so many times, it doesn't work that
> way anymore.
>
> >   What was 'the
> > industry standard' prior to HSpice?
> >=3D20
> That was probably Berkeley SPICE, and/or a handful of close
derivatives.
> But unless I'm mistaken, fewer people were running SPICE back then.
>
> But sooner or later it all comes down to model creation.  You can't
get
> good SPICE results without good models, and if the experts who create
> the device-specific models for their FAB line, need to use better MOS
> models than those that come with Berkeley SPICE 2g6 or even 3f5, then
> they move on to something newer.
>
> > -- What's attractive about HSpice vs. the competition?
> >=3D20
> The models my vendor supplies are HSPICE proprietary, so I have little
> if any choice.
>
> W-line models are nice.
>
> Regards,
> Andy
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
> List archives are viewable at:
>                 //www.freelists.org/archives/si-list
> or at our remote archives:
>                 http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
>                 http://www.qsl.net/wb6tpu
>

------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:    =20
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages=20
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
 =20
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:    =20
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages=20
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
 =20

------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
                //www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu
  

Other related posts: