Subject: Re: Dynamic/Global Parameters in Verilog-AMS
From: Kevin Cameron x3251 (Kevin.Cameron@nsc.com)
Date: Wed Feb 28 2001 - 17:41:09 PST
> From oleary@cadence.com Wed Feb 28 10:30:41 2001
>
> Kevin,
>
> I don't think adding a new parameter type is such a major extension.
> After all the 1364-2001 supports parameter, specparam and localparam.
> Also I have been careful to define the dynamicparam so that it only
> effects features that have been added for AMS.
The semantics (rather than the syntax) looked like a major extension
to me.
> As you pointed out using `define does not allow sweeping. I don't totally
> understand your solution to this. What did you mean by sweeping ranges on
> parameters and using attributes somehow. Have you an example of what you mean?
You said wanted the ability to alter parameters, and that that functionality was
for sweeping in different analyses. The general attribute scheme should allow you
to add attributes indicating that a parameter is sweepable and whether the
change should propagate, ranges for parameters are already permitted. Rather
than add 'dynamicparam', you could do something like:
module res(p,n); inout p,n; electrical p,n;
parameter real (* dynamic,
sigma=0.1, distribution="gaussian",
layer="M2" *) r=0 from (0:inf];
analog V(p,n) <+ r*I(p,n);
endmodule;
- indicating the parameter 'r' is sweepable and its tolerance.
[ Note: I didn't have the (*...*) syntax reference handy, so I'm guessing.]
Your system (and anyone else's who wants to implement it) can treat the
parameter as you would 'dynamicparam', but other tools can just ignore
the attribute.
Kev.
> --Martin
> On Feb 27, 9:36am, Kevin Cameron x3251 wrote:
> > Subject: Re: Dynamic/Global Parameters in Verilog-AMS
> >
> > 'dynamicparam' seems like a major extension to me, I think you can use
> `define to
> > create "global" definitions that would do part of what your asking for, and
> we
> > already have ranges on parameters which could be used for sweeping etc. -
> attributes
> > can tell the tools how to use them.
> >
> > Is it really necessary?
> >
> > Kev.
> >
> >-- End of excerpt from Kevin Cameron x3251
>
>
>
>
This archive was generated by hypermail 2b28 : Fri Mar 02 2001 - 17:08:29 PST