RE: [SI-LIST] : modeling languages (was: receiver jitter)

About this list Date view Thread view Subject view Author view

From: Muranyi, Arpad ([email protected])
Date: Fri Jan 21 2000 - 11:59:36 PST

Kumar, and everyone,

I would also add the .MEASURE statements to your list in #2.

I am using HSPICE, and their .MEASURE statements are pretty good
compared with other SPICE flavors (in some of which it doesn't exist).
Yet there are a lot of measurement types that you can't do, or just very
painfully. For example: try measuring a non-monotonic signal crossing a
certain voltage threshold region for the last time for each edge of a
multi-cycle simulation. Painful! And the list could go on...

Arpad Muranyi
Intel Corporation

-----Original Message-----
From: C. Kumar [mailto:[email protected]]
Sent: Friday, January 21, 2000 10:33 AM
To: [email protected]
Subject: Re: [SI-LIST] : modeling languages (was: receiver jitter)


There seems to be lots of misunderstanding and
confusion over spice, modeling etc. Let me add my
comments. Hopefully it will not add to the confusion.

1. IBIS should remain as a data standard. It should
not be going in the direction of "node description
language" because one already exists viz spice. It is
powerful, flexible and time tested. No good will come
out of reinventing the wheel. IBIS should be focusing
on things like adding thing time-current curves which
will aid greatly SSN modeling and resolve the silly
ambiguties inherent in time-voltage curves.

Fir die/pkg/connectors the langauage to store and pull
in appropriate subcircuits should be standardized at
the cad tool level.

2. Some crucial element types in spice need to be

For starters a IBIS behavior block element needs to be

Coupled lossy transmission line elements

A class of controlled current/voltage sources needs to
be standardized. These sources lie at the heart of a
circuit simulator and the present selection available
in most spice simulators is both unncessarily varied
(like CCVS CCCS etc etc) and yet insuffficient.

3. If item 2 is addressed anything can be modeled.
period. Things like receiver models become merely
standardized subcircuits which will show the same
behavior in different simulators.

**** 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

About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Apr 20 2000 - 11:34:47 PDT