[ibis-macro] Re: Problem with IBIS-AMI clock_times definition
- From: Mike Steinberger <msteinb@xxxxxxxxxx>
- To: Arpad_Muranyi@xxxxxxxxxx
- Date: Thu, 18 Jun 2009 14:27:49 -0500
Arpad-
Responses below.
Mike S.
Muranyi, Arpad wrote:
Mike,
Thanks, this makes perfect sense. So going back to the
original question, shouldn't we say something about this
in the IBIS-AMI spec?
I see this as being a bit like teaching my grandmother to suck eggs;
however if people believe this would be helpful as a programming
suggestion in the spec, fine, let's do it. I do suggest that we not make
a big deal of this because it's not a big deal.
I don't think we can "prescribe" and enforce what algorithms
people should/can put into the DLL, but at least give them
a ***strong*** suggestion (and example) for how one might do
it properly to avoid problems with multi million bit runs.
"Here's one way to do it that we know works." is about as far as I would
go. Furthermore, I don't think it would need much more text than my
previous e-mail.
In addition, may I request that you update your tool kit
with this in mind (preferably with a 64-bit integer) to set
a correct example for the novice model makers?
The technique that's in there is already not too bad in that the number
of floating point operations is one plus the number of previous GetWave
calls; however, it could be improved to use exactly the technique I
described. The source code changes would be simple enough, but I'd be a
bit more concerned about the effort to post a new version. Would the
benefit to novice model makers really justify that effort? That's not
obvious to me.
Just a technical detail question on the "bit twiddling" topic.
This may not always be necessary, depending on what language
the model maker is using, right? I would assume that more
modern languages would allow someone to use 64-bit integers
in multiplication with a floating point variable without
additional user intervention. Or am I wrong?
Among the languages I've used- C/C++, PERL, Verilog, Visual Basic,
MatLab/Octave, python- Verilog has a "time" variable type that's 64
bits. My recollection of MatLab/Octave is a bit hazy, so maybe there's
something there, though I don't think so. Any of these languages on a 64
bit machine probably also uses 64 bit integers. Other than that, I truly
do mean to leave 64 bits as an exercise for the student.
Thanks,
Arpad
==============================================================
-----Original Message-----
From: Mike Steinberger [mailto:msteinb@xxxxxxxxxx]
Sent: Thursday, June 18, 2009 1:22 PM
To: Muranyi, Arpad
Cc: IBIS-ATM
Subject: Re: [ibis-macro] Re: Problem with IBIS-AMI clock_times definition
Keep the sample count as an integer, then use a single floating point
multiplication to multiply by the sample interval to generate each
floating point time value. The integer arithmetic has no round-off
error, and there's only one floating point operation. Mirabile dictu and
Bob's your uncle.
If 32 bits isn't large enough to count all the samples you have in mind,
then create a double precision integer (e.g., 64 bits), increment that,
and perform the floating point multiplication. That means you have to
write some bit twiddling functions to perform the arithmetic functions.
I think those bit twiddling functions would make a good homework
assignment for an introductory programming course, since it would
require an understanding of binary arithmetic, provide practice with
bit-wise logical operations, and force a deeper understanding of the
relationship between integer and floating point variables in the
programming language.
Mike S.
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
Other related posts:
