[ewg] FW: Suggestions for Characterizing Transition Times for Infiniband

[Rupert Dance]

Sorry all - this went to the wrong EWG - should have been IBTA

 

Rupert

 

From: ewg-bounces at lists.openfabrics.org
[mailto:ewg-bounces at lists.openfabrics.org] On Behalf Of Rupert Dance
Sent: Wednesday, April 02, 2014 11:17 AM
To: ewg at lists.openfabrics.org
Cc: 'Tranchemontagne, Michael P'; Alexander Rysin; Hiroshi Goto;
bob_hasenick at non.agilent.com; John Smith; James Morgante; Joe O'Brien;
john.c.calvin at tektronix.com; OJ Danzy
Subject: [ewg] FW: Suggestions for Characterizing Transition Times for
Infiniband

 

Hi All,

 

As you know, we are conducting the IBTA PF25 for the next two weeks and we
have in attendance many engineers from the Test Equipment vendors:

 

Agilent

.         Bob Hasenick

.         OJ Danzy

 

Anritsu

.         James Morgante

.         Hiroshi Goto

 

Tektronix

.         John Calvin

.         John Smith

.         Mike Tranchemontagne

 

Wilder Technologies

.         Joe O'Brien

 

There are several topics we would like to discuss today:

 

1.      Patterns and conditions used to characterize transition times for
the InfiniBand testing - see discussion below <> 

 

2.      Use of Clock Recovery during ATD Testing 

 

3.      Ideal length(s) for Wilder HCBs

 

Thanks

 

Rupert

 

From: bob_hasenick at non.agilent.com [mailto:bob_hasenick at non.agilent.com] 
Sent: Wednesday, April 02, 2014 10:30 AM
To: rsdance at soft-forge.com
Cc: James.Morgante at anritsu.com; john.c.calvin at tek.com
Subject: Suggestions for Characterizing Transition Times for Infiniband

 

Hi Rupert:

 

                We discussed the need to be more specific about the patterns
and conditions used to characterize transition times for the Infiniband
testing we're doing at the Plugfest 25.  Currently the standard says to use
PRBS9, and the test is performed on the eye.  The reported values can vary
greatly depending on several factors.  Other standards have addressed this
by being more specific in one of two ways.  I assume that the needs to
characterize transition times within Infiniband are very similar to the
needs of other standards so the Infiniband method could be similar.

 

One:  Use the first 18 bits of a PRBS9 pattern and measure rise time on two
specific transitions using the waveform rather than the eye.  An example of
this is shown from OIF CEI below.

 

Two:  Use a pattern with repeating number of zeros and ones such as square 4
or square 8.  An example of this is shown from IEEE below, and is included
in the CEI reference.

 

                Please let me know as you have questions.

 

Thanks,

Bob

 

 

 



 



 

Bob Hasenick

Technical and Applications Consulting

707-577-3425



 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20140402/243c11a1/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: application/octet-stream
Size: 46556 bytes
Desc: not available
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20140402/243c11a1/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.jpg
Type: application/octet-stream
Size: 30390 bytes
Desc: not available
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20140402/243c11a1/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.jpg
Type: application/octet-stream
Size: 19510 bytes
Desc: not available
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20140402/243c11a1/attachment-0002.obj>