Re: [IBIS] RE: [ibis-macro] Re: BIRD107: Update to Algorithmic Modeling API (AMI) Support in IBIS

From: C. Kumar <kumarchi_at_.....>
Date: Fri Apr 04 2008 - 04:46:24 PDT
I strongly recommend that the wave form should be generated ONLY by the eda tool under all conditions. The Tx and Rx model ONLY MODIFY, the waveform. i.e they filter the waveform and should not be in the business of generating waveform per se

On a side note there is no stipulation that the waveform can only generated by convolution.  There is absolutely no reason for such a requirement. 

This is particularly true when GetWave call is involved. The impulse response in the Init can be construed as an APPROXIMATION to the channel.

AMI interface is a DATA FLOW interface and is not concerned with how the wave form is generated. 

When the EDA tool is only guy in town in the business of generating the waveform, it can and should decide how it is done and the particular way it is done has no bearing on the AMI interface. That is the original intent of the interface.

Without such a clear and justifiable division of labor there will be a lot of confusion an doubt. Besides it is the right thing!!!


Todd Westerhoff <twesterh@sisoft.com> wrote: Arpad,

Now I see what you're saying.  I think that this case works as well, but the evaluation order may
not be what you were thinking.

It's late here - I'd like to defer this one until later this morning, so that I've got some time to
think through this & speak with others.

Okay?

Todd. 


Todd Westerhoff
VP, Software Products
SiSoft
6 Clock Tower Place, Suite 250
Maynard, MA 01754
(978) 461-0449 x24
twesterh@sisoft.com
www.sisoft.com

-----Original Message-----
From: ibis-macro-bounce@freelists.org [mailto:ibis-macro-bounce@freelists.org] On Behalf Of Muranyi,
Arpad
Sent: Friday, April 04, 2008 12:23 AM
To: ibis@eda.org; IBIS-ATM
Subject: [ibis-macro] Re: BIRD107: Update to Algorithmic Modeling API (AMI) Support in IBIS

Todd,

I think we don't understand each other.

What I was referring is this.  Your explanation for that change in
BIRD107
was that for the FALSE case you couldn't require to have the EDA tool
convolve the unchanged channel impulse response with the stimulus
waveform
pattern (as I suggested it one of my emails) because if this was
required,
AND there was an RX model which also had a FALSE then the channel
impulse
response would be convolved into the total outcome twice (= double
counted).

My most recent question was suggesting that this could also happen with
what is in BIRD107 for the TRUE case.  In the TRUE case the output of
INIT
is convolved with the stimulus waveform and sent into the GETWAVE.  If
this
is done for both the Tx and Rx, then the modified impulse response
coming
out of INIT will be included twice in the total calculations.  (This is
with the assumption that the INIT function's input is the channel
impulse
response for Tx and Rx).

Is this situation never going to happen?

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

-----Original Message-----
From: ibis-macro-bounce@freelists.org
[mailto:ibis-macro-bounce@freelists.org] On Behalf Of Todd Westerhoff
Sent: Thursday, April 03, 2008 9:02 PM
To: ibis@eda.org; 'IBIS-ATM'
Subject: [ibis-macro] Re: BIRD107: Update to Algorithmic Modeling API
(AMI) Support in IBIS

Arpad,

I think the effect is the same.

TX: Use_Init_Output=False  RX: Use_Init_Output=False
TX: Use_Init_Output=True   RX: Use_Init_Output=False

In either case, the output from the TX AMI_Getwave should be fed
directly into the RX AMI_Getwave.

Agree?

Todd.

Todd Westerhoff
VP, Software Products
SiSoft
6 Clock Tower Place, Suite 250
Maynard, MA 01754
(978) 461-0449 x24
twesterh@sisoft.com
www.sisoft.com
-----Original Message-----
From: ibis-macro-bounce@freelists.org
[mailto:ibis-macro-bounce@freelists.org] On Behalf Of Muranyi,
Arpad
Sent: Thursday, April 03, 2008 11:23 PM
To: ibis@eda.org; IBIS-ATM
Subject: [ibis-macro] Re: BIRD107: Update to Algorithmic Modeling API
(AMI) Support in IBIS

That double counting could also happen when you have a Tx GetWave and
an Rx GetWave call in the same system with the Boolean being FALSE
both times.  Or is that never going to happen?

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


-----Original Message-----
From: ibis-macro-bounce@freelists.org
[mailto:ibis-macro-bounce@freelists.org] On Behalf Of Todd Westerhoff
Sent: Thursday, April 03, 2008 8:14 PM
To: ibis@eda.org; 'IBIS-ATM'
Subject: [ibis-macro] Re: [IBIS] BIRD107: Update to Algorithmic Modeling
API (AMI) Support in IBIS

Oops.

Off by one word.  Corrected copy below.

Arpad,

I made that change when preparing the BIRD.  The presentation we were
reviewing on Tuesday was
focusing solely on a TX model how the Init/Getwave calls would be used
to predict the waveform at
the RX pad.

What I realized while preparing the BIRD was that the language we were
using could be construed as
requiring a second copy of the channel impulse response to be convolved
into the waveform presented
to the RX AMI_Getwave call, even when Use_Init_Output is False.

Consider the case where the TX and RX model both implement filtering in
AMI_Init and AMI_Getwave.
In the case where Use_Init_Output for the RX model is set to False, then
the output of the TX
AMI_Getwave call should be fed directly into the RX AMI_Getwave call.

The existing language

| If the Reserved Parameter, Use_Init_Output, is set to "False", EDA
tools will
| use the original (unfiltered) impulse response of the channel.

Didn't support that clearly.  My concern was that the impulse response
of the channel would end up
in the output twice in some EDA implementations.  The proposed language
was ambiguous in the case
where both models have both calls, and Use_Init_Output is False.

This was a realization I had at 5:00 this afternoon, with different key
people out of the office.
Thus I had a decision to make - change the language and go ahead (I know
this stuff works, we've
already tested and correlated it) - or wait.  I chose the former.

Sorry if this came as a surprise - there's no hidden agenda here.  There
was simply a case we forgot
to consider, and I made a change to address it.

We can discuss this further at tomorrow's and next Tuesday's meeting.

Todd.

Todd Westerhoff
VP, Software Products
SiSoft
6 Clock Tower Place, Suite 250
Maynard, MA 01754
(978) 461-0449 x24
twesterh@sisoft.com
www.sisoft.com

-----Original Message-----
From: owner-ibis@eda.org [mailto:owner-ibis@eda.org] On Behalf Of
Muranyi, Arpad
Sent: Thursday, April 03, 2008 8:26 PM
To: ibis@eda.org; IBIS-ATM
Subject: RE: [IBIS] BIRD107: Update to Algorithmic Modeling API (AMI)
Support in IBIS

Excuse me, but how did the language on pg. 18 of this presentation:
http://www.vhdl.org/pub/ibis/macromodel_wip/archive/20080401/toddwesterh
off/IBIS-AMI%20Correlation%20and%20BIRD%20Update/IBIS_ATM_BIRD_Update_04
0108.pdf
change from:

| If the Reserved Parameter, Use_Init_Output, is set to "False", EDA
tools will
| use the original (unfiltered) impulse response of the channel.
| The algorithmic model is expected to modify the waveform in place.

into what is in the text of the BIRD107:

| If Use_Init_Output is set to "False", the EDA platform will present
the 
| input waveform directly to the AMI_Getwave call (i.e. without
convolving 
| the waveform with the impulse response returned by AMI_Init).

The two paragraphs have a significant difference and I didn't see
any discussion on that change...  Could someone please explain that
to me?  What happened with the "impulse response of the channel"?

Thanks,

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

-----Original Message-----
From: owner-ibis@server.eda.org [mailto:owner-ibis@server.eda.org] On
Behalf Of Mirmak, Michael
Sent: Thursday, April 03, 2008 5:09 PM
To: ibis@server.eda.org
Subject: [IBIS] BIRD107: Update to Algorithmic Modeling API (AMI)
Support in IBIS


The enclosed BIRD, "Update to Algorithmic Modeling API (AMI) Support in
IBIS," is submitted on behalf of Todd Westerhoff, SiSoft and Zhen Mu,
Cadence Design Systems.  It will be introduced and discussed at an
upcoming IBIS Open Forum teleconference.

All resolved and pending BIRDs can be found at http:// <>
www.eda.org/ibis/birds/.

- Michael Mirmak
  Intel Corp.
  Chair, IBIS Open Forum

 

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


--------------------------------------------------------------------
|For help or to subscribe/unsubscribe, e-mail majordomo@eda-stds.org
|with the appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       
|  subscribe   ibis-users 
|  unsubscribe ibis       
|  unsubscribe ibis-users 
|
|or e-mail a request to ibis-request@eda-stds.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda-stds.org/pub/ibis/email_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/users_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/email/         E-mail since 1993

---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  http://www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@freelists.org
  Subject: unsubscribe

---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  http://www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@freelists.org
  Subject: unsubscribe


---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  http://www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@freelists.org
  Subject: unsubscribe

---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  http://www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@freelists.org
  Subject: unsubscribe



-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

--------------------------------------------------------------------
|For help or to subscribe/unsubscribe, e-mail majordomo@eda-stds.org
|with the appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       
|  subscribe   ibis-users 
|  unsubscribe ibis       
|  unsubscribe ibis-users 
|
|or e-mail a request to ibis-request@eda-stds.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda-stds.org/pub/ibis/email_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/users_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/email/         E-mail since 1993


       
---------------------------------
You rock. That's why Blockbuster's offering you one month of Blockbuster Total Access, No Cost.
-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


--------------------------------------------------------------------
|For help or to subscribe/unsubscribe, e-mail majordomo@eda-stds.org
|with the appropriate command message(s) in the body:
|
|  help
|  subscribe   ibis       <optional e-mail address, if different>
|  subscribe   ibis-users <optional e-mail address, if different>
|  unsubscribe ibis       <optional e-mail address, if different>
|  unsubscribe ibis-users <optional e-mail address, if different>
|
|or e-mail a request to ibis-request@eda-stds.org.
|
|IBIS reflector archives exist under:
|
|  http://www.eda-stds.org/pub/ibis/email_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/users_archive/ Recent
|  http://www.eda-stds.org/pub/ibis/email/         E-mail since 1993
Received on Fri Apr 4 04:54:01 2008

This archive was generated by hypermail 2.1.8 : Fri Apr 04 2008 - 04:54:46 PDT