RE: Using 'E' as the extrapolation character in table_model

I agree with Geoffrey's statement. The 'E' extrapolation character is supposed to indicate a region where the model cannot be trusted anymore so no data from that region should be used. Extrapolation into this forbidden region should be treated similar to square root of a negative number or divide-by-zero.

Cheers,
Marq

-----Original Message-----
From: owner-verilog-ams@eda.org [mailto:owner-verilog-ams@eda.org] On Behalf Of Geoffrey Coram
Sent: Friday, April 29, 2011 5:03 PM
To: Dave Miller
Cc: Verilog-AMS LRM Committee
Subject: Re: Using 'E' as the extrapolation character in table_model

I would say it should be treated similar to a math error --
the simulator should not use any value that is returned.
The simulator should then get a chance to shrink the
timestep or whatever it usually does when it gets a math
error from device evaluation.

Dave Miller wrote:
> I have a question that was raised recently from Cary Renzema (refer mantis 3527)
>
> If I use 'E' as the extrapolation character in $table_model() what should the
> behaviour be?
> 'E' means we error, but exactly what sort of error.
>
> Is it a hard error - do we stop simulation right there.
>
> Or do we only raise the error at the end of the time step, and if so what value
> should we return from the table_model. Should it be clamped to the closest
> endpoint, or should we use linear extrapolation.
>
> I guess we should detail what happens when using 'E' to remove the ambiguity.
>
> Anyone have comments on what should happen here?
>

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