Re: Minutes of the Verilog-AMS meeting: 24th July 2008

From: Paul Floyd <Paul_Floyd_at_.....>
Date: Fri Jul 25 2008 - 09:02:20 PDT
Geoffrey.Coram wrote:

Hi all

> expr_tol also!

> In this case, I don't know what a zero tolerance would mean.
>
> Are we going to say that only positive tolerances are allowed
> (since we haven't defined behavior for zero), or are we going
> to say "non-negative" and just leave out a description of
> what zero means?  Or say the behavior for zero is undefined
> and/or simulator dependent?

I agree for expr_tol, but there is a bit of a difference between the 
expression and time tolerances. The expression is effectively continuous 
(down to 1 unit in the last place [ULP], about 1 part in 1e17).  This 
governs the effective minimum meaningful expression tolerance.

The time is discrete, with a simulator-dependent step size (and equally 
dependent minimum step size).  So for the time tolerance, the smallest 
meaningful value is this minimum step size.

In either case, I think that if a value of 0 is to be accepted, should 
behave the same as if the tolerance were ommitted. Now that 
cross/above/timer can have null arguments, I don't see that there is any 
real benefit, other than a lazy way of saying "give me the smallest 
possible tolerance".

Regards
Paul Floyd
-- 
Dr Paul Floyd
Mentor Graphics Corporation


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Fri Jul 25 09:02:48 2008

This archive was generated by hypermail 2.1.8 : Fri Jul 25 2008 - 09:02:51 PDT