[ibis-macro] Re: AMI Info Model_Specific parameters: Market Speed vs Standards Speed
- From: "Walter Katz" <wkatz@xxxxxxxxxx>
- To: "'Scott McMorrow'" <scott@xxxxxxxxxxxxx>
- Date: Thu, 24 Mar 2011 17:26:30 -0400 (EDT)
Scott,
I agree it is not "standard" until IBIS approves it. Both we and our IC
Vendors and System Designer customers understand this. But we have a
problem to solve and we are trying to solve it in as open a way as
possible. This is the commitment we have made to our IC Vendor partners
and to our System Designer customers. We had growing pains with OpalT, but
I point out that we have brought to IBIS Backchannel and Repeaters very
early in the cycle.
Walter
From: Scott McMorrow [mailto:scott@xxxxxxxxxxxxx]
Sent: Thursday, March 24, 2011 5:14 PM
To: Walter Katz
Cc: ibis-macro@xxxxxxxxxxxxx
Subject: Re: [ibis-macro] Re: AMI Info Model_Specific parameters: Market
Speed vs Standards Speed
I'll will retract the word "sneak" and modify my comments as follows:
So essentially these mechanisms are being used create a technical and
competitive market advantage with new IBIS_AMI non-standard features,
without going through the standardization process first. This allows the
model to be "advertised" as IBIS-AMI compliant. Afterwards, once the
implementation is fait accompli, a model maker comes back to the committee
to standardize what has become a de facto standard for that model.
The EDA vendor whose tool was used for development of the initial
non-standard IBIS-AMI model that complies with the standard (because the
parameters passed are just "information") gains certain market advantages.
The silicon vendor also gains a technical market advantage by enabling new
features that are fully exploited by one EDA tool. Documentation of the
"parameters" is most likely provided in a document delivered with the
IBIS-AMI models, describing how the "parameters" are to be used, but ...
the EDA vendor whose tool was used for development of the initial
non-standard standard model has already implemented automation of the
documented "parameters", such that this is the only EDA tool that the
initial models will have full operational capability in. That's all well
and good. As long as we all understand and accept that the features will
eventually be standardized in due time. (Although it would be really nice
for this to occur before the press releases come out.)
From my perspective an informational parameter that is being used to alter
modeling or simulation is not standard, and should not be advertised as
standard, until it has been turned into a reserved parameter and
documented. There are too many cases of model makers introducing features
to solve one problem, while ignoring (or even creating) others.
respectfully
Scott
Scott McMorrow
Teraspeed Consulting Group LLC
121 North River Drive
Narragansett, RI 02882
(401) 284-1827 Business
(401) 284-1840 Fax
http://www.teraspeed.com
TeraspeedR is the registered service mark of
Teraspeed Consulting Group LLC
On 3/24/2011 4:53 PM, Walter Katz wrote:
Scott,
We have a problem to solve, we did not sneak this in. We published
everything we have done and entertained all reasonable requests for
change. I do not think you can ask us to do anything else.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Scott McMorrow
Sent: Thursday, March 24, 2011 4:43 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: AMI Info Model_Specific parameters: Market Speed
vs Standards Speed
Walter
So essentially these mechanisms are being used to sneak in new IBIS_AMI
features without going through the standardization process first. And
this allows the model to be "advertised" as IBIS-AMI compliant.
Afterwards, once the implementation is fait accompli, a model maker comes
back to the committee to standardize what has become a de facto standard
for that model. Is my understanding correct?
Scott
Scott McMorrow
Teraspeed Consulting Group LLC
121 North River Drive
Narragansett, RI 02882
(401) 284-1827 Business
(401) 284-1840 Fax
http://www.teraspeed.com
TeraspeedR is the registered service mark of
Teraspeed Consulting Group LLC
On 3/24/2011 4:30 PM, Walter Katz wrote:
Arpad,
The fundamental problem that we face is the need for IC Vendors to supply
System Companies with models that they can use today to design next
generation high speed products. Info parameters in BIRDS 121-124 are
required to do this today. The market cannot wait for IBIS to approve
this, and then wait for EDA vendors to support it. It is critical that
when it is required to add information to a .ami file that a simulator
must use (e.g. a Model_Specific Info parameter), that the parameter be
published by submitting a BIRD to make it a Reserved_Parameters. The need
for this was described in the OpalT document, and in numerous IBIS-ATM
discussions. In the meantime EDA Vendors have the information they need to
support these advanced capabilities as their markets require them. This is
exactly what Sigrity and Gennum have done with their backchannel proposed
BIRD, and SiSoft will determine when the specification of backchannel is
firm enough, and then support it in our tool when our customers require
it.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Thursday, March 24, 2011 2:54 PM
To: IBIS-ATM
Subject: [ibis-macro] Re: Table Clarification BIRD
If I remember correctly, we spent a considerable time
on this topic in the last ATM teleconference. In that
discussion I think I heard Walter state something
along the lines of who cares, the EDA tool just copies
everything to the parameter string for DLL and doesn't
need to know whether the first column is a row number
or useful data.
Then the question was raised, how about (Usage Info)
in which case the table is to be consumed by the tool,
and in that case the tool would have to know what the
first column is.
This brought us to the topic of syntax consistency and
ease of parsing, etc.
But this also raises another issue, namely whether the
Model_Specific parameters should ever be (Usage Info).
Note that this is not a problem with Reserved_Parameters,
which supposed to be well defined in the spec.
Comments?
Arpad
==========================================================
Walter Katz
wkatz@xxxxxxxxxx
Phone 303.449-2308
Mobile 720.333-1107
Other related posts:
