From: Fred Balistreri ([email protected])
Date: Mon Jan 24 2000 - 14:29:29 PST
The encription in Hspice is not for the scaling but to protect the
IC vendors topology and transistor models...ie process. However
as mentioned by DC, the encription can cause you problems for
SI analysis since once encripted you cannot change naming, scaling
or anything.
Scaling is also not good if your getting your models from various
vendors. Since some vendors don't use universal scaling, some do,
some have localized scaling etc. Its true one can always "flatten"
the netlist at the expense of time.
Some earlier versions of Hspice on the PC only had limited support of
the CMOS devices. If you had that version of Hspice you were out of
luck unless you upgraded.
Again the primary reason for Hspice usage is IC. Today there are two
compeling reasons (in my mind anyway) to use Hspice.
1. Wide industry acceptance and support.
2. CMOS models.
However as mentioned before Hspice is not necesarlity a great choice
for SI work (W element aside).
Best Regards,
Chris Cheng wrote:
> there is no reason why the scaling factor needs to be encrypted.
> besides, hspice can have different scaling factor for different
> transistor model.
> chris
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of D. C. Sessions
> Sent: Monday, January 24, 2000 7:44 AM
> To: [email protected]
> Subject: Re: [SI-LIST] : modeling languages (was: receiver jitter)
>
> Jim Freeman wrote:
> >
> > Hi Fred,
> > Hspice will accomodate different models from different vendors but not
> automatically. If scale is the same for different models, all you
> > have to do is rename the models from other vendors with a unique name and
> edit the netlist to reference the same name. I have simulated up to
> > 8 different technologies in the same simulation using this technique. I
> have no idea how high it can go but I suspect the limit has to do more
> > with the simulation machine than the Hspice code.
>
> That's fine if you have editable models.
> The problem arises when people propose *encrypted* models as
> a solution to the distribution-of-proprietary-data problem.
>
> > Fred Balistreri wrote:
> >
> > > Hello Jim, the answer is no. Not even Hspice will do that if the
> > > parts are from different vendors that use different modeling
> > > methodologies. Hspice also uses propietary models, those
> > > cannot be done by law. However the levels that are public
> > > are compatible. That is one of the reasons why I said I
> > > understand why people use Hspice.
> > >
> > > However realize that Hspice is designed for IC not SI
> > > engineers. Where as Aspice is designed for SI engineers.
> > > The advantages we have are not in the IC areas but rather
> > > in the SI area. I realize that transistor level model do play
> > > an important role and for that we do have a solution although
> > > it may not be as "clean" as one would like.
> > >
> > > Best Regards,
> > >
> > > Jim Freeman wrote:
> > >
> > >> Part 1.1 Type: Plain Text (text/plain)
> > >> Encoding: 7bit
> > >
> > > --
> > > Fred Balistreri
> > > [email protected]
> > >
> > > http://www.apsimtech.com
> > >
> > >
>
> --
> D. C. Sessions
> [email protected]
>
> **** To unsubscribe from si-list: send e-mail to
> [email protected]. In the BODY of message put: UNSUBSCRIBE
> si-list, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>
> **** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
-- Fred Balistreri [email protected]**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu ****
This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:34:49 PDT