Re: [sv-dc] 4-state vs 3-D

To go back to the start: "4-state" doesn't really apply to something like "real-valued" signals or user-defined signals, so I was trying to break "4-state" down into its constituent parts to make it easier to discuss with respect to other types. I.e. the 4-state signal value is really a degenerate case of the more general 3-D case where the dimensions are:

   Strength
   Certainty
   Value

So 4-state is:

   Z = zero-strength (or undriven)
   X = non-zero strength (driven), but we are uncertain whether it's 1 or 0
   0/1 = driven and the value is known

Dimensionally - bool,bool,bool.

For resolving the value of the net you first ask what's the highest strength driver (first dimension), does it have a known value (second dimension), and then what is the value (third dimension). A simulator can handle the 3-D representation any way you like, but to the programmer it just looks like a four value enum.

So for any discrete type you might want to ask what values does it carry on any of the dimensions (none, enum, real, or class/struct). The second question is then how to do the type conversions where dimensions collapse. Using NaN appears as an option when collapsing the uncertainty dimension and the value dimension of a "real valued" type.

Kev.

On 12/10/2010 09:17 AM, Lear, Jim wrote:
> Sure, but the subject states "4-state vs 3-D." Are there two proposals or concepts we're trying to weigh, such as enumerated vs structure?
>
> From the electrical domain 'Z' is reasonably straightforward. The 'X' is a new concept for that side, but the NaN is quite similar. The subtleties are escaping me.
>
> Kindest Regards,
> Jim Lear
> Cirrus Logic
> (512) 851-4612
> (512) 293-7248 (mobile)
>
> -----Original Message-----
> From: Scott Cranston [mailto:cranston@cadence.com]
> Sent: Friday, December 10, 2010 10:28 AM
> To: Lear, Jim; 'sv-dc@eda.org'
> Subject: RE: [sv-dc] 4-state vs 3-D
>
> I think we're trying to decide what is being modeled in real modeling and (a subdecision) what X means in that model.
>
> -- Scott
>
>
> -----Original Message-----
> From: owner-sv-dc@eda.org [mailto:owner-sv-dc@eda.org] On Behalf Of Lear, Jim
> Sent: Friday, December 10, 2010 9:25 AM
> To: 'sv-dc@eda.org'
> Subject: RE: [sv-dc] 4-state vs 3-D
>
> I'm not sure how important this is, but in the AMS world with boundary models (connect rules), an 'X' can also represent a Voltage that is between the zero- and one-threshold; the meta-stability region. This arises in the conversion from a known Voltage to a logic value.
>
> I'm not sure exactly what we're trying to solve. Are we trying to decide between a "four state" and a "3-D" implementation? It's not clear what exactly is meant between the two. Would someone be kind enough to clarify?
>
> Kindest Regards,
> Jim Lear
> Cirrus Logic
> (512) 851-4612
> (512) 293-7248 (mobile)
>
>
>

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.