RE: [IBIS] *-AMS modeling and hurting companies?

From: Mike LaBonte <milabont_at_.....>
Date: Fri Apr 01 2005 - 22:02:26 PST
Few will disagree that AMS has more potential than SPICE. Arpad gives good
examples of advanced chip design features that programming-style languages
(AMS) lend themselves to well. At the other extreme, IBIS 2.1 is fine for a
great majority of parts. In retrospect, maybe the "middle ground" of
*somewhat* advanced models (equalization, etc.) has actually been served
well enough by a combination of HSPICE and design kits written using the
proprietary macromodel language of each simulator. So maybe macromodel
extensions to IBIS are not needed so badly while waiting for AMS to ramp up.
Maybe I am underestimating what it would take for IBIS to adopt what seems
like a few relatively simple changes to implement macromodeling.

I'm not strongly on either the AMS or macromodel bandwagons, per se. I'm
happy just working with the IBIS Quality committee helping IBIS 2.1 files to
serve us well. My suspicion is that the simulator companies who have IBIS
capability but not AMS are also doing fine with their favorite advanced
SPICE with IBIS 2.1 support and macromodel-type extensions, and they feel
that there is plenty of time to get ready for AMS. But I would like to hear
from them. The future will bring changes and no doubt Arpad is already
experimenting with those changes, as usual :)

Mike

-----Original Message-----
From: owner-ibis@eda.org [mailto:owner-ibis@eda.org] On Behalf Of Muranyi,
Arpad
Sent: Friday, April 01, 2005 1:41 PM
To: ibis
Subject: [IBIS] *-AMS modeling and hurting companies?

Mike,

I would like to respond to two of your earlier points.

1)  "As someone with a software background I would love to play with AMS
languages, but having been successful with the macromodel approach all along
I see no need."

This depends greatly on what you want to model.  I doubt that you will be
able to make (efficient) behavioral models for the upcoming fancy buffer
technologies, such as ones having FIR filters, decision feedback loops,
clock recovery circuits, etc...  I am afraid that if we don't prepare the
capabilities for these now, we will again lag behind with our modeling
capabilities when the need is burning.
At that time people will ***again*** continue bashing IBIS saying:
see, IBIS can never keep up with design technology, and there is nothing
else that works but SPICE.

I agree, *-AMS may be a lot more than what we need today for most of our
modeling needs, but in order to get out of this lagging mode I think we need
to be proactive and start getting familiar with *-AMS now, so that by the
time we will really need it we should have something in place that is well
proven widely implemented by tool vendors, and is familiar to the industry.
It will be too late to start scratching our head then in attempt to find a
viable solution.


2)  "Also, as far as I know AMS involves a lot of buck$ and would leave all
but 2 simulator companies with a big disadvantage."


Let's just look at some facts.  Here is the top portion of BIRD75.8, the
language extension BIRD to IBIS:

BIRD ID#:        75.8
ISSUE TITLE:     Multi-Lingual Model Support
REQUESTER:       Bob Ross and Chris Reid, Mentor Graphics, Arpad Muranyi
                 and Michael Mirmak, Intel DATE SUBMITTED:  3/29/02, 5/3/02,
7/15/02, 8/14/02, 9/11/02, 9/27/02,
                 10/18/02, 12/20/02, 12/23/02 DATE ACCEPTED BY IBIS OPEN
FORUM: 1/10/03


Notice the submission dates, they span about ten months or so.

Also, the IBIS 4.1 specification including this BIRD was ratified on January
30, 2004, more than a year after the last version of BIRD75.8 was submitted.
This indicates that the idea of the *-AMS language extensions was openly
discussed for about two years before it was voted to become part of the
official IBIS specification.  It didn't come as a surprise.  There was
enough time to find out when the big vote will occur for IBIS 4.1.

Quote from the minutes of the January 10, 2003 IBIS Open Forum
teleconference in which the last version of this BIRD was voted on:

"With no further discussion Stephen Peters called for a vote on both
BIRD77.2 and BIRD75.8 together.  Both BIRDs were approved unanimously."

Quote from the minutes of the January 30, 2004 IBIS Open Forum
teleconference:

"Bob Ross moved to vote on approval of the draft specification as IBIS 4.1.
The motion was seconded by John Angulo.  Michael called for a vote, and the
draft document was unanimously approved as IBIS 4.1."

Also, I counted 29 and 30 companies under the "VOTING MEMBERS AND 2003
PARTICIPANTS LIST" and "VOTING MEMBERS AND 2004 PARTICIPANTS" in the above
mentioned two meeting minutes.  This represents more than just those two
simulator companies you are referring to in your message.  I wonder, how
come none of them showed up, or spoke up in these meetings that this would
hurt them?  How come we are still getting these kinds of comments now,
exactly three years after the first version of this BIRD, BIRD75.1 was
officially submitted?

By the way, the very first presentation in IBIS meetings on VHDL-AMS and
Verilog-AMS was done at DAC2000 on June 8, 2000, almost five years
ago:  http://www.eda.org/pub/ibis/summits/jun00/bakalar.zip

Let me ask, what is going on here?  None of the so-called hurting companies
showed up to vote, and we should refrain from releasing
*-AMS models now because we need to feel sorry for them?   Hmmm, I
don't get this picture...

Arpad
======================================================================

-----------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org with the 
|appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       <optional e-mail address, if different>
|  subscribe   ibis-users <optional e-mail address, if different>
|  unsubscribe ibis       <optional e-mail address, if different>
|  unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda.org/pub/ibis/email_archive/  Recent  
| http://www.eda.org/pub/ibis/users_archive/  Recent
|  http://www.eda.org/pub/ibis/email/          E-mail since 1993
-----------------------------------------------------------------
|For help or to subscribe/unsubscribe, email majordomo@eda.org
|with the appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       <optional e-mail address, if different>
|  subscribe   ibis-users <optional e-mail address, if different>
|  unsubscribe ibis       <optional e-mail address, if different>
|  unsubscribe ibis-users <optional e-mail address, if different>
|
|or email a request to ibis-request@eda.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda.org/pub/ibis/email_archive/  Recent
|  http://www.eda.org/pub/ibis/users_archive/  Recent
|  http://www.eda.org/pub/ibis/email/          E-mail since 1993
Received on Fri Apr 1 22:02:33 2005

This archive was generated by hypermail 2.1.8 : Fri Apr 01 2005 - 22:03:21 PST