[ibis-macro] Re: FW: some suggestion for AMI proposal
- From: Huang chunxing <huangchunxing@xxxxxxxxxx>
- To: ckumar@xxxxxxxxxxx, "Ward, Richard" <richard.ward@xxxxxx>
- Date: Thu, 23 Nov 2006 09:37:50 +0800
Kumar,
Yes, that is my meaning.
Richard,
If it is true, then I take back my suggestion. : )
I may misunderstand the ami proposal. In the slide, I only find equalizers
models. In my thoughts, the only possible way to handle FEC is in DFE equalizer
model for initial ami proposal. The FEC model will be within DFE model and
different chip vendors may give different FEC models . But I think FEC is a
normal process and it will be better that is done by EDA vendor. The
calculation only needs dfe coefficients. So, I suggest api return the dfe tap
value.
Regards,
Huang
Huawei Technologies Co.,Ltd.
Tel: 86 755 28976229
FAX: 86 755 28976758
Email: huangchunxing@xxxxxxxxxx
Web: http://www.huawei.com
----- Original Message -----
From: Ward, Richard
To: ckumar@xxxxxxxxxxx ; Huang chunxing
Cc: ibis-macro@xxxxxxxxxxxxx
Sent: Wednesday, November 22, 2006 1:47 AM
Subject: RE: [ibis-macro] Re: FW: some suggestion for AMI proposal
Kumar/Huang,
Given any FEC will just work on blocks of (equalized) data, it
would seem sensible to incorporate that logic in the algorithmic
model (if required)- but I can't see that any changes to the
current API proposal are needed to accomdate it.
As for the DFE error propagation figure- this is based off the
steady-state tap values and is a relatively simple spreadsheet
style calculation, so I believe the same applies (all the data
required for the calculations are already available inside the
DLL/whatever and therefore no extra IO is required (although
mabe I'm missing parametric control of the new functions?).
Regards,
Richard
------------------------------------------------------------------------------
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of C. Kumar
Sent: 21 November 2006 05:26
To: Huang chunxing
Cc: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: FW: some suggestion for AMI proposal
so just to clarify, you want 'error correction methods' as part of the
initial standard. am i reading your right?
----------------------------------------------------------------------------
From: Huang chunxing [mailto:huangchunxing@xxxxxxxxxx]
Sent: Tuesday, November 21, 2006 7:39 AM
To: C. Kumar
Cc: ibis-macro@xxxxxxxxxxxxx
Subject: Re: [ibis-macro] FW: some suggestion for AMI proposal
Kumar,
Thanks for your reply. I really wish initial version of AMI could have
solution for analytical model of error propagation and FEC. IBIS model could
handle this analytical model just as receiver equalizer.
10Gbps standards for backplane have FEC available. CEIP of OIF have
enclosed fire code with ability of correcting 7 burst error. 802.3ap has
optional QC(2112,2080) to correct 11 burst error. So I believe in order to
extend IBIS to 10Gbps and even higher data rate, analytical model of error
propagation and FEC is just as important as receiver equalizer and it will be
better for you to finish this work at this time.
Regards,
Huang
Huawei Technologies Co.,Ltd.
Tel: 86 755 28976229
FAX: 86 755 28976758
Email: huangchunxing@xxxxxxxxxx
Web: http://www.huawei.com
----- Original Message -----
From: Lance Wang
To: ibis-macro@xxxxxxxxxxxxx
Cc: C. Kumar ; Ray Komow ; Hemant Shah ; Ken Willis
Sent: Monday, November 20, 2006 11:16 PM
Subject: [ibis-macro] FW: some suggestion for AMI proposal
--------------------------------------------------------------------------
From: C. Kumar
Sent: Monday, November 20, 2006 9:26 AM
To: Lance Wang
Cc: Hemant Shah; Ken Willis; Ray Komow
Subject: RE: [ibis-macro] some suggestion for AMI proposal
lance, can you send this on behalf of me?
huang:
Thanx for your comments. I have heard and read about the importance of
dfe error propagation.
This is first version of the spec (AMI) which is confined to basic
filtering of signals. The interface functionally separates the filtering, from
other data processing. So strictly it addresses physical or virtual
modification of a data stream.The data stream is generated and handed over by
the EDA domain, which is then filtered and modified in the ip domain.
The focus of this current ami interface is to keep it simple yet useful
and evolvable.
While this version of the interface indeed allows for passing unlimited
amount of filter specific parameters, it does not contain knowledge and
actions pertaining to particular filter types like DFE. e.g. It does not
specifically address issues like error propagation.
Your suggestions will be most welcome.
------------------------------------------------------------------------
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Huang chunxing
Sent: Monday, November 20, 2006 4:28 AM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] some suggestion for AMI proposal
Hi Kumar and all,
First let me introduce myself. My name is Huang Chunxing and I'm a
backplane engineer from Huawei.
I have read your AMI proposal to IBIS macro model task group. It's a
great idea to solve that IBIS couldn't handle the DSP algorithm problems. In
the paper, I noticed that AMI also need to meet the requirement of proprietary
post-processing of data. I totally agree that.
In my opinion, the error propagation of DFE is very good case for this.
One error bit of DFE will effectively change the error probability of
sequential bits, depending on the DFE taps and coefficients. If we want to
analyse the DFE error propagation effect, we need to know the coefficients of
DFE. But I didn't find any output params of the defined ami functions related
to it, including AMI_init and AMI_getwave. In order to keep the IBIS with most
updated DSP algorithm, some extra user defined params may be required. I hope
you and group member will consider my suggestion. By the way, I know little
about dll. If my understanding is wrong, I apologize first. Thanks!
Regards,
Huang
Huawei Technologies Co.,Ltd.
Tel: 86 755 28976229
FAX: 86 755 28976758
Email: huangchunxing@xxxxxxxxxx
Web: http://www.huawei.com
Warning: The information contained in electronic mail message is
intended
only for the personal and confidential use of the designated
recipient(s) named
above. It may be privileged and confidential. If you have received
this communication
in error, please destroy any and all copies of this message
including attached
files in your possession.
Other related posts:
