[ibis-macro] Re: EMD Question
- From: <rrwolff@xxxxxxxxxx>
- To: <Arpad_Muranyi@xxxxxxxxxx>
- Date: Mon, 14 Jul 2008 16:10:32 -0600
Arpad,
Here are answers to your poll and can be considered Micron's current
opinion on EMD.
We struggle with modeling of packages and memory DIMMs using the current
IBIS specification. EBD does not provide a very useful model when it
comes to DDR3 technology and beyond. We end up providing some customers
with SPICE formatted netlists for specific signals on our memory
modules, but this model format is difficult to create, support, and use.
When modeling multi-die packages, the EBD format still has limited
support and is not adequate for modeling loss characteristics. The IBIS
package model options for modeling single-die packages are woefully
inadequate as well.
I liked the idea of developing a standard set of accepted SPICE-like
elements for EMD, but I place more importance on getting a modeling
standard developed quickly. I'm sure accepted standards will quickly
develop based on individual EDA company support, but I would like to see
some emphasis placed on getting most of the EDA companies to agree on
some acceptable formats.
Thanks,
Randy
1) Do you want IBIS-ATM to pursue this interconnect
modeling methodology?
Yes.
2) Do you need this type of package, cable and connector
interconnect model?
Yes.
3) What do you dislike about this EMD?
I am currently concerned with not knowing what languages will be
supported by EDA vendors. I would eventually like to see some
official responses from EDA companies as to what languages will be
supported.
4) What do you need added to EMD?
The concept of "views" is not mentioned in the summary below, but I
think this is a very important component of EMD. Like [Model Selector]
for component models, "views" would be very useful for supplying
multiple formats of models with varying levels of complexity, all within
the same EMD wrapper.
5) Would you deliver models for packages and connectors in
this format?
I would specifically like to deliver models of packages and memory DIMMs
in this format.
6) Would you use models for packages and connectors in this
format?
Yes, assuming my EDA software fully supports it.
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Tuesday, July 08, 2008 6:10 PM
To: IBIS-ATM
Subject: [ibis-macro] EMD Question
Hello everyone,
We had an interesting (and long) IBIS-ATM meeting today, and at the end
we decided to poll the membership of this email list to find out whether
we are going in the right direction with EMD.
To summarize, EMD, as currently proposed, is kind of a generalization of
EBD. It is like EBD in that it will contain a [Pin List] section that
describes the external connections of a module, and a [Reference
Designator Map] that describes instances of IBIS components and other
EMD modules. [Extended Nets] are a new feature in EMD that is not
directly in EBD but is equivalent to the "Pins"
and "Nodes" within a single EBD [Path Description].
Unlike EBD, EMD allows [Extended Nets] between Nodes without being
connected to a Pin. [Path Description]s get replaced with subckt calls
to models that can be written in one or more simulation languages.
These subckts will model interconnect between Pins and Nodes, or between
nodes. In addition, subckt calls can have ports with Pins and Nodes
from two or more [Extended Nets], and can also have ports to supply nets
on either component instances or pins. The "free form" EMD discussed
today explicitly does not constrain the languages or subset of languages
that subckts are written in other than the content of these subckts are
simulation models that describe the interconnect between Pins and Nodes.
Here are a few questions we would like to get answers from you to make
sure we are doing the right thing:
1) Do you want IBIS-ATM to pursue this interconnect
modeling methodology?
2) Do you need this type of package, cable and connector
interconnect model?
3) What do you dislike about this EMD?
4) What do you need added to EMD?
5) Would you deliver models for packages and connectors in
this format?
6) Would you use models for packages and connectors in this
format?
Thanks,
Arpad
===============================================================
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
Other related posts:
