From: Gregory R Edlund (firstname.lastname@example.org)
Date: Mon Dec 04 2000 - 13:34:04 PST
I am one of the folks who worked on the I/O Buffer Accuracy Handbook
(formerly IBIS Accuracy Specification). Let me try to address your
1. I don't know how widely used our work is, but I have talked to a
handful of people who are using it. It's hard to say what most people are
doing with it; the sample size isn't very large!
2. I have written a C program that reads two waveform files and computes
the figure of merit we documented in the Handbook. It also does two other
computations since the first one can be misleading in certain cases. I
just secured permission from my company to publish this code as freeware.
If you can wait a few days, I'm planning to have it on the IBIS web site
next week (if our web master is available). Of course, visual checks are
always helpful, too. We still haven't built a computer that comes close to
the human brain!
3. So far I have not seen this procedure used in any simulation tools,
although I've talked to three different EDA people who have shown an
interest in pursuing it. Perhaps you should call your local rep! :) One
hang-up we found in trying to store a "golden waveform" in an IBIS file is
that the simulator actually tries to use the data in constructing its
stimulus waveform. Ugly simulations result. What we really need is some
syntactical construct in IBIS that says, "This is a golden waveform." I've
been thinking about authoring a BIRD (IBIS buffer issue resolution
document) to get this syntax into IBIS. (Too many irons in the fire.) We
would also need a modification to s2ibis2 that creates these golden
waveforms using the new IBIS syntax. These things would go a long way
toward improving the accuracy of model data, eh?
IBM Server Technology Development
3605 Hwy. 52 N, Dept. HDC
Rochester, MN 55901
Date: Fri, 1 Dec 2000 11:45:40 -0400
From: "Brad Crowell" <email@example.com>
Subject: [SI-LIST] : Verifying IBIS models
I am wondering how the verification of an IBIS model is being handled. The
accuracy spec provides the "golden waveforms" that a simulation output can
be checked against, but how is this checking being done by most people? Is
this just a visual check or are there some tools out there that can provide
a quantitative measure of the model's accuracy? Is this feature included in
some of the simulation tools?
Thanks for your comments,
77 Chain Lake Drive
Halifax, Nova Scotia, B3S 1E1
Tel: (902) 450-1700 ext#287
Fax: (902) 450-1704
firstname.lastname@example.org. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
This archive was generated by hypermail 2b29 : Tue May 08 2001 - 14:30:20 PDT