[ibis-interconn] Re: Suggested Plan for IBIS Packaging and On_Die Interconnect Modeling
- From: "Muranyi, Arpad" <Arpad_Muranyi@xxxxxxxxxx>
- To: Ibis-Interconn <ibis-interconn@xxxxxxxxxxxxx>
- Date: Fri, 21 Sep 2012 15:29:24 +0000
Lynne,
BIRD 125 does just that. It adds a new keyword to the [Define Package Model]
keyword to allow IBIS-ISS subcircuits to be used in place of the RLC matrices.
This has the benefits of adding all that is possible in IBIS-ISS (T-stone files,
W-elements, and all the passive elements R, L, C, K) to the IBIS package models.
It would also remove the one-to-one mapping of pins to pads.
But this doesn't provide a solution for the desire of going "buffer centric"
in IBIS. As you know, IBIS is "Component (or Pin) centric", meaning that you
start with a Component name, instantiate the package, and the package is
connected between pins and assumed or named pads, and the buffer model is also
instantiated by the pin list. For pre layout simulation people would like to
have models without going by a pin list, because the chip design may not be done
and may not have a pin list yet, while buffer models and package models for
certain specific channels or interfaces might already exist.
BIRD 125 currently doesn't have a solution for stacked die packages or MCM-s
either, partly because IBIS doesn't have a way to define the dice inside a
Component.
Solving some of these deficiencies would involve a more major overhaul of IBIS.
BIRD 125 doesn't address EBD, but a similar BIRD could be written to do the
same to our current EBD spec to add IBIS-ISS to EBD.
As far as impact on model makers go, the impression I am getting from the
discussions is that nobody is using the package keywords of IBIS and whether
we implement BIRD 125 or come up with a brand new syntax, the model makers
would have to learn it from scratch. While BIRD 125 would have minimal impact
on the parser and EDA tools, it is not welcomed with open arms because it
doesn't provide a solution for all of the desires people have for IBIS, and
some people feel that its syntax (based on the original package syntax) is ugly.
I hope this answers your questions. Thanks,
Arpad
================================================================================
-----Original Message-----
From: ibis-interconn-bounce@xxxxxxxxxxxxx
[mailto:ibis-interconn-bounce@xxxxxxxxxxxxx] On Behalf Of Lynne C. Green
Sent: Thursday, September 20, 2012 8:42 PM
To: Ibis-Interconn
Subject: [ibis-interconn] Re: Suggested Plan for IBIS Packaging and On_Die
Interconnect Modeling
Since I am new to this group, I would like to pose a Question to the group:
What would members think of starting with a "quick" solution of adding
one keyword for language (Touchstone, ISS-SPICE) to [Package Model]?
Benefits? Limitations?
Some preliminary thoughts along those lines:
- Fast implementation, with a minimum of new keywords, no new model type.
- Might also work for EBD?
- Could be made flexible enough for future needs (Walter's keywords)?
- Impact on EDA tools (person-months)?
- Impact on parser development?
- Impact on model makers, EDA tools, model users?
- Could on-die power distribution info be captured in [Model]?
Personally, I am philosophically inclined to keep package and die
separate within IBIS. Often a chip can go in any of several packages,
and a package can fit any of several chips.
Cheers,
Lynne
"IBIS training when you need it, where you need it."
Dr. Lynne Green
Green Streak Programs
http://www.greenstreakprograms.com
425-788-0412
lgreen22@xxxxxxxxxxxxxx
------------------------------------------------------------------
The IBIS Ad Hoc Interconnect Task Group Mailing List
Archives are available at:
//www.freelists.org/archives/ibis-interconn
TO UNSUBSCRIBE:
Send a message to "ibis-interconn-request@xxxxxxxxxxxxx"
with a subject of "unsubscribe"
To administer your subscription status from the web, visit:
//www.freelists.org/list/ibis-interconn
Meeting minutes and files are available; visit:
http://www.eda.org/ibis/interconnect_wip/
------------------------------------------------------------------
The IBIS Ad Hoc Interconnect Task Group Mailing List
Archives are available at:
//www.freelists.org/archives/ibis-interconn
TO UNSUBSCRIBE:
Send a message to "ibis-interconn-request@xxxxxxxxxxxxx"
with a subject of "unsubscribe"
To administer your subscription status from the web, visit:
//www.freelists.org/list/ibis-interconn
Meeting minutes and files are available; visit:
http://www.eda.org/ibis/interconnect_wip/
Other related posts:
