[SI-LIST] Re: package SSN model accuracy requirements
- From: "Mirmak, Michael" <michael.mirmak@xxxxxxxxx>
- To: <si-list@xxxxxxxxxxxxx>
- Date: Mon, 14 Mar 2005 14:33:49 -0800
Chris et al,
Here are some responses to your questions. Please note that I cannot
speak for my employer as a whole; my company is large and different
divisions with diverse product lines have customers with differing
needs. I can only speak for what I have observed.
a) Does your company believe these so-called secondary effects like
package parasitics and SSO will impact your I/O performance ?
In a word, absolutely! We expend considerable effort to optimize
package and power system designs for best I/O performance.
b) Does your company feels it is necessary to provide a model that can
model and analyze these effects correctly ?
Again, unquestionably yes. Detailed driver and package models are used
internally and also provided to our customers. Moreover, just as Ray
stated in his response, we are working on a variety of efforts to
improve correlation between internal and customer models to actual lab
performance. =20
c) What model is provided TODAY to address those analysis needs ?
The answer is, effectively, "all kinds." We provide IBIS as well as
just about every proprietary tool format, included encrypted ones,
though not every division provides the same kinds for every product.
Customers, particularly in different market segments, have incredibly
diverse requirements. We are, in my experience, responsive to customer
feedback -- their existing tools and experience have a big impact on
their modeling requests -- and also try to demonstrate model accuracy
directly to our customers.
d) What CAD tool is currently supporting those models so that your
customers can analyze the problem accurately ?
=20
I would simply say "see above." If a particular division is
distributing a proprietary format as a result of large customer demand,
then that tool's features -- good or bad -- are the only ones available
to those customers. More specifically, we too are researching a large
number of modeling formats and styles for improving our support of SSN
and other power integrity effects. Ideally, we will unify behind a
single modeling format. When we can demonstrate concrete results, from
lab and correlation to our internal design format's output, we will
offer models in that format.
As you can see, my answers do not differ much from Ray's. I do want
make some additional comments. For any modeling solution -- including
those used for power integrity analysis -- we've found seven basic
demands from interested parties:
"I want it to be accurate"
"I want it to be fast in my simulator"
"I want it to protect my IP"=09
"I want it to be standard" (works for more than one tool)
"I want it available soon" (wait for standard or tool upgrade?)
"I want it easy to use/implement/automate"
"I want maximum flexibility in describing my design's behavior"
These include the needs of the model author, model user and tool vendor.
Not every customer will have all of these desires, nor have them in the
same priority order. I do not believe any single solution can hit more
than six of these at any one time. However, the refrain we consistently
hear from the industry -- borne out by the vendor poll data from the
DesignCon IBIS Summit -- is that speed of simulation is near the top of
most lists. I think the winning solution for the industry, in the short
and long terms, will satisfy as many of the above as possible.=20
On IBIS and SSN: many in the industry recognize that, for effects such
as SSN, traditional IBIS (3.2, 4.0) today does not include enough data
to provide a consistent power integrity result between tools. As only
at-IO-pad V-T data is provided, tools have to estimate how the power and
ground rail currents are distributed. The assumptions made will
determine the SSO effects seen. The BIRD95 efforts demonstrated at the
recent DesignCon and DATE summits show that adding additional data
specifically on rail currents can achieve better correlation to the
original design while keeping simulation speed high.
- Michael Mirmak
Intel Corp.
Chair, EIA IBIS Open Forum
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
