[ibis-macro] Re: ignore_bits and TX vs. RX
- From: "Walter Katz" <wkatz@xxxxxxxxxx>
- To: <Arpad_Muranyi@xxxxxxxxxx>, "'IBIS-ATM'" <ibis-macro@xxxxxxxxxxxxx>
- Date: Wed, 27 Jun 2012 18:11:51 -0400 (EDT)
Arpad,
Mostly academic. I do know of some 16 and 32 bit Tx FFE equalizers that
are coming down the road, but I do not think that there will ever be a
need for a Tx to have Ignore Bits >~30. Any Rx that has a DFE takes many
bits to equalize, often ~500K. I do not know of an Rx that does not have
some sort of dynamic equalization (either DFE, FFE, or more complex
equalization). But some words to the affect that the Rx equalization
algorithm is suspect until after the Tx Ignore Bits occurs would be OK.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Wednesday, June 27, 2012 5:49 PM
To: 'IBIS-ATM'
Subject: [ibis-macro] Re: ignore_bits and TX vs. RX
Walter,
Maybe in the case of a small vs. large number in the
Tx and Rx this may not make a difference, but what if
both Tx and Rx have large numbers, like 500,000 in each?
If the user asks for 1 million bits, they could get
nothing if we add them up and if the tool doesn't do
what I described in my previous reply.
I wonder if the spec should say something about this,
as a suggestion (not requirement).
Thanks,
Arpad
=========================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Walter Katz
Sent: Wednesday, June 27, 2012 4:42 PM
To: Muranyi, Arpad; 'IBIS-ATM'
Subject: [ibis-macro] Re: ignore_bits and TX vs. RX
Arpad,
Technically the EDA tool should probably ignore the sum of the two ignore
bits. This is most likely academic since Ignore_Bits for a Tx is usually
on the order of 3, while for Rx it is usually 50K or sometimes 500K.
Walter
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Wednesday, June 27, 2012 5:15 PM
To: IBIS-ATM (ibis-macro@xxxxxxxxxxxxx)
Subject: [ibis-macro] Re: ignore_bits and TX vs. RX
Mike,
I tend to agree with Ambrish. The EDA tool should take
the larger of the two numbers, and exclude the first so
many bits from being plotted.
I may add, though that the tool could also do other things.
For example, if the user requests 1 million bits to be
simulated, the tool could simulate 1000035 bits and plot
the last 1 million bits for the user (after ignoring the
first 35).
The spec doesn't say how this should be handled.
Thanks,
Arpad
=============================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Ambrish Varma
Sent: Wednesday, June 27, 2012 3:44 PM
To: michael.mirmak@xxxxxxxxx; IBIS-ATM (ibis-macro@xxxxxxxxxxxxx)
Subject: [ibis-macro] Re: ignore_bits and TX vs. RX
Michael,
I would say that the tool would need to ignore only 35 bits in the
waveform for deducing the final results. My understanding is that the tool
does not really 'shorten' the number of bits by the ignore bits, but
actually simply 'ignores' it in the final analysis for results etc, as
that is the number of bits required for the models to adapt/optimize or
settle down.
Thanks,
Ambrish.
Ambrish Varma | Member of Consulting Staff
P: 978.262.6431 <http://www.cadence.com> www.cadence.com
_____
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Mirmak, Michael
Sent: Wednesday, June 27, 2012 4:31 PM
To: IBIS-ATM (ibis-macro@xxxxxxxxxxxxx)
Subject: [ibis-macro] ignore_bits and TX vs. RX
A deceptively simple question.
Ignore_bits as a parameter can apply to the TX or the RX. The language of
the specification states that it determines the number of bits to be
ignored by the simulation tool.
In the case where I have a link where only an isolated TX or an isolated
RX, this is easy to understand. But what happens when I combine a TX with
an RX on the same link where both use ignore_bits?
I would assume that the TX (using GetWave) would provide the tool with a
waveform that the tool would then shorten by the stated number of bits
(say, 21. But would the RX GetWave then use the shortened waveform in its
own analysis, before the tool reports back an even shorter waveform?
Specifically, if ignore_bits is 21 in the TX, and ignore_bits is 35 in the
RX, does that mean that the final waveform reported out by the tool would
be shorter than the original pattern by 56 bits?
- MM
Other related posts:
