Re: [ng-spice] Who is Mike Seningen?
Paolo Nenzi wrote:
>
> On Tue, 13 Jul 1999, Mike Seningen wrote:
>
> > That will be very useful... I know a little spanish :-),
> > I believe Italians close :-)
>
> > Third: To make sure that everything is clean. Please
> > let me know when I have said something or aluded to something
> > which could expose your code to information that could be
> > detrimental.
> I am sorry, i do not understand, what do you mean ?
>
You have the gist -- I want to make sure that my exposure
to the tools and their terminology does not cause problems.
But I think from you follow-up things will be fine.
> > That having been said:
> >
> > A couple of areas that I am most familiar with are:
> >
> > Measurement:
> > All the commercial tools have various means for measuring
> > data. I've seen some do it post-processing, but most
> > do it internal to the run, some even use this information
> > for future iterations.
> Are you talking about a program like the Silvaco's UTMOST ?
My understanding of UTMOST is they use it to characterize
VLSI processes, like a arbitrary 0.15u process as an example.
What I was referring to was the ability to say something like
measure delay in transient mode from node a at a pt 1 to
node b at pt 2, etc etc.
These measures would be part of your deck and you could
use it for informative reasons or use it as program control
"delay" could be some variable in a parameter. Much like
.Modif you mentioned below.
>
> >
> > The second is the output. My understanding is that there
> > is only a textual output which allows you to graph in
> > ASCII, or Tabular formats... All comercial tools use
> > some sort of binary output. Hspice is the defacto std. (the .tr0
> > format) Smartspice has emulated this as a .raw file.
> > I do not know if there is a actual standard used.
> I think that most commercial tools uses binary output to compact the files
> size and to discourage from take the results and import them in another
> program.
> >
Exactly....
> > A third area is in the use of parameters. It is my understanding that a
> > berkely spice deck requires hard coded
> > numbers. I assume the commercial releases use a pre-processor
> > to take in parameterized netlists and internally convert to
> > a hard coded deck.
> Yes, spice require hard coded parameters and that's bad. There is a perl
> script thet partially overcome the problem, but an extension to the parser
> is needed.
yes, most commercial tools consider Hspice compatability to
be the "standard" for parsing purposes.
> >
> > A fourth exploits the ability to internally measure and
> > the use of parameters. Different commercial versions
> > name it differently, and they appear to handle slightly
> > differently but in general its called "Optimization".
> > It uses various Numerical Iterative Analysis methods
> > to solve for some result by being able to vary certain
> > given parameters. This could be classified as a frill,
> > but it is a very powerful frill.
>
> Yeah, I would like to see a .MODIF on Berkeley's Spice.
Bingo -- but I didn't want to use that term :-)
Does Berkley have .alter? sweeps?
>
> > A fifth feature we use a lot (but I do not know if
> > it is in the Berkely spice) is the Voltage/Current
> > controlled sources.
> Spice3f4/5 is able to simulate all the four types of controlled sources,
> plus an analog behavioral model. The analog behavioral model is an
> arbitrary function I/V source. A sort of simple "a" device of SmartSpice.
>
Also correct -- I have used all of he above.
> > The last feature I know about is usually various
> > proprietary spice model tweaks that are built into
> > the programs. Obviously this would not be something
> > you need to support as long as you support Bsim3v3.
> Yes we support bsim3v3 and in plan is: BSIM3SOI.
Great :-)
>
> Now some news, by the end of the week there will be a tarball ready for
> testing. I will put the tarball on the ftp archive:
> ftp://ieee.ing.uniroma1.it/pub/electronics
>
> Plase note this is a really alpha version but it will support:
>
> 1) Bsim3v3.2.2
> 2) Temp sweep (to be tested) by Serban Popescu.
>
Jeez -- sounds like you don't need me :-)
Seriously, obviously I have some catching up to do,
maybe private email of your status and where/what I can
do specifically to help. Or I can just monitor things...
> Bye,
> Paolo
>
> PS: How is the temp. there in Houston ?
Well from I'm actually in Austin, about 150miles WNW of
Houston, Temp has been quite mild this year -- we're
low to mid 90's. Houston's about the same just more
Humid (it's closer to the Gulf Coast).
--
---
Michael Seningen
EVSX, Inc.
11612 RM 2244 Suite 200
Austin, Texas 78733
512/421-2134 Fax:512/263-0795
email: mikes@evsx.com
---
Partial thread listing:
- Re: [ng-spice] Who is Mike Seningen?, (continued)