[ibis] Re: BIRD 178.2 recommendation from the ATM group

  • From: <radek_biernacki@xxxxxxxxxxxx>
  • To: <wkatz@xxxxxxxxxx>, <tom@xxxxxxxxxxxxxxxxx>, <Arpad_Muranyi@xxxxxxxxxx>, <ibis@xxxxxxxxxxxxx>
  • Date: Tue, 28 Jul 2015 00:10:05 +0000

Hi Tom,

Yes, what you mention is a valid solution, however outside of the IBIS spec
and the question raised by Arpad.

Thanks,
Radek

From: Walter Katz [mailto:wkatz@xxxxxxxxxx]
Sent: Monday, July 27, 2015 5:06 PM
To: Tom Dagostino; BIERNACKI,RADEK (K-Sonoma,ex1); Arpad_Muranyi@xxxxxxxxxx;
ibis@xxxxxxxxxxxxx
Subject: RE: [ibis] Re: BIRD 178.2 recommendation from the ATM group

Tom, Radek, Arpad,

The EDA tool can do lots of things, but let's just leave that up to the EDA
tool. IBIS 6.0 is deficient here, but we do not need to come up with scenarios
on how to hand IBIS 6.0 models that are I/O. All the user would need to do is
set his IBIS version to 6.1, and the .ami version to 6.1, and modify the IBIS
file to change the Executable statements in the I/O models. To Exectuable_Rx
and Executable_Tx as required.

Remember, the standard requires that any IBIS 6.0 file and before can be
changes to IBIS 6.1 and all will work (except for Use_init_Output).

Walter

From: Tom Dagostino [mailto:tom@xxxxxxxxxxxxxxxxx]
Sent: Monday, July 27, 2015 8:00 PM
To: radek_biernacki@xxxxxxxxxxxx<mailto:radek_biernacki@xxxxxxxxxxxx>;
wkatz@xxxxxxxxxx<mailto:wkatz@xxxxxxxxxx>;
Arpad_Muranyi@xxxxxxxxxx<mailto:Arpad_Muranyi@xxxxxxxxxx>;
ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Subject: RE: [ibis] Re: BIRD 178.2 recommendation from the ATM group

Radek

Erroring out seems very unfriendly, why not open a dialogue with the user and
ask what is desired?

Tom Dagostino

Teraspeed Labs
9999 SW Wilshire Street
Suite 102
Portland, OR 97225

971-279-5325
tom@xxxxxxxxxxxxxxxxx<mailto:tom@xxxxxxxxxxxxxxxxx>

From: ibis-bounce@xxxxxxxxxxxxx<mailto:ibis-bounce@xxxxxxxxxxxxx>
[mailto:ibis-bounce@xxxxxxxxxxxxx] On Behalf Of
radek_biernacki@xxxxxxxxxxxx<mailto:radek_biernacki@xxxxxxxxxxxx>
Sent: Monday, July 27, 2015 3:30 PM
To: wkatz@xxxxxxxxxx<mailto:wkatz@xxxxxxxxxx>;
Arpad_Muranyi@xxxxxxxxxx<mailto:Arpad_Muranyi@xxxxxxxxxx>;
ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Subject: [ibis] Re: BIRD 178.2 recommendation from the ATM group

Hi Arpad,

The user can mix/match files with different IBIS versions. As far as BIRD 178.2
is concerned it is very clear and in the case of an I/O models, the model
"Enable" will define the direction (although only one direction may be
supported, if defined as such by the model maker). Any older versions of IBIS
files should be interpreted the same way as they are today.


Executable is prohibited if the Model_type for the associated [Model] is "I/O",
"I/O_open_drain", "I/O_open_sink", "I/O_open_source", or "I/O_ECL".
Executable_Tx, Executable_Rx:
The Executable_Tx and Executable_Rx subparameters are alternatives to the
Executable subparameter, for I/O-capable buffers. The arguments (fields)
supported are syntactically identical to the Executable subparameter. At least
one Executable_Tx or one Executable_Rx subparameter is required if the
Model_type for the associated [Model] is "I/O", "I/O_open_drain",
"I/O_open_sink", "I/O_open_source", or "I/O_ECL". For all other Model_types
where [Algorithmic Model] is supported, only the Executable subparameter is
permitted. In these cases, the direction of the associated [Algorithmic
Model]s shall be assumed by the EDA tool to follow the [Model] Model_type
declaration.


It is not to the EDA tool to assume any "opposite" direction for the "other"
model. The EDA tool should rather error out if no Rx or no Tx is present.

Radek

From: ibis-bounce@xxxxxxxxxxxxx<mailto:ibis-bounce@xxxxxxxxxxxxx>
[mailto:ibis-bounce@xxxxxxxxxxxxx] On Behalf Of Walter Katz
Sent: Monday, July 27, 2015 11:34 AM
To: Arpad_Muranyi@xxxxxxxxxx<mailto:Arpad_Muranyi@xxxxxxxxxx>;
ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Subject: [ibis] Re: BIRD 178.2 recommendation from the ATM group

Arpad,

I am not sure what you mean by "uses the features".

If you mean an Algorithmic Model section in an I/O [Model], then the problem is
not defined to the EDA tool.

If you mean that a channel has an AMI model in an Output [Model], and an AMI
model in an I/O [Model], then one could argue that the AMI model in the I/O
[Model] is an Rx model. This becomes problematic if that same AMI model in an
I/O [Model] is used in a channel with an AMI model in an Input [Model].

If you mean what happens if a .ami file has Rx Reserved Parameters and is uses
in an Output [Model], then the Rx parameters should be ignored.

Or did you mean something else?

Walter

From: ibis-bounce@xxxxxxxxxxxxx<mailto:ibis-bounce@xxxxxxxxxxxxx>
[mailto:ibis-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Monday, July 27, 2015 2:13 PM
To: ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Subject: [ibis] Re: BIRD 178.2 recommendation from the ATM group

I would like to ask a question.

What is the EDA tool expected to do when there are two AMI models
in a channel, one which uses the features outlined in this BIRD,
and another one which is an older model without these features.

Does the assumption that the older model without these parameters
must be the "opposite type" apply in this case? If so, should
this be mentioned in the BIRD? (Or is this there already, did
I just miss it)?

Thanks,

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



From: Bob Ross [mailto:bob@xxxxxxxxxxxxxxxxx]
Sent: Thursday, July 23, 2015 8:52 PM
To: ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Cc: Muranyi, Arpad
Subject: RE: [ibis] BIRD 178.2 recommendation from the ATM group

IBIS reflector,

The pending editorial requests for the teleconference meeting are
documented below, and the BIRD178.2 with approved amendments would
become BIRD178.3.

The minor requested editorial changes involve terminology in Table 1:

Change Any to Rx, Tx
Change Rx-only to Rx
Change Tx-only to Tx

Also, after the end of the ANY OTHER BACKGROUND INFORMATION section and
before the first sentence: "Add the following text at the end of Section 10.7,
...", add

"For each reserved parameter add one of these Direction choices, as documented
in Table 1
below, with the format positioned before the Descriptors heading:


Direction: Rx

Direction: Tx

Direction: Rx, Tx

Bob

From: ibis-bounce@xxxxxxxxxxxxx<mailto:ibis-bounce@xxxxxxxxxxxxx>
[mailto:ibis-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Tuesday, July 21, 2015 1:02 PM
To: ibis@xxxxxxxxxxxxx<mailto:ibis@xxxxxxxxxxxxx>
Subject: [ibis] BIRD 178.2 recommendation from the ATM group

Hello IBIS Folks,

BIRD 178.2 "Specifying Buffer Directionality for AMI" was discussed
in today's IBIS-ATM teleconference, and a vote was taken to make a
recommendation to the IBIS Open Forum to schedule a vote on it to
approve it (with a minor editorial change) in the next IBIS Open
Forum teleconference with the goal to include the BIRD in the upcoming
IBIS specification which is currently being worked on in the IBIS
Editorial Task Group.

This email serves the purpose of officially relaying this decision
and recommendation to the IBIS Open Forum.

Questions or comments will be addressed in the IBIS Open Forum
teleconference in which this BIRD is scheduled to be voted on.

Thanks,

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

Other related posts: