RE: VHDL-AMS PLI and a Question about the Simulation Cycle

Sameer,

The LRM does indeed specify the interaction between the analog time step algorithm and the
event-driven engine, but only to the extent of defining the quality of the results that
are generated. This was done to obtain, in the limit of infinite precision, a portable
solution. The LRM does not define how to obtain the results with the specified properties,
leaving it to an implementor to choose any suitable algorithm. For competitive reasons I
cannot give you any such information either, you have to come up with your own solution.

Ernst

On Thu, 4 Jan 2007 17:41:39 -0500, Sameer Kher wrote:
> Thanks for taking the time to respond Ernst.
>
> Perhaps my question was not properly framed. I do understand that once
> determined, the analog solution point for a cycle need not be
> re-determined unless required and also that the analog solver solves at
> the start of every cycle with no direct dependence on digital events.
>
> I am trying to determine if a particular synchronization strategy is
> specified by the language for the analog and digital simulators (by
> allowing the analog solver to only advance at most by the next digital
> event time).
>
> A simple mixed signal example - a 2 bit counter with the most
> significant bit affecting an analog system which has a time varying
> solution. Per the LRM, won't the analog solver be executed even when the
> least significant bit changes?
>
> Thanks,
> Sameer
>
>
> -----Original Message-----
> From: Ernst Christen [mailto:Ernst.Christen@synopsys.COM]
> Sent: Thursday, January 04, 2007 3:55 PM
> To: Sameer Kher
> Cc: vhdl-ams@vhdl.org
> Subject: RE: VHDL-AMS PLI and a Question about the Simulation Cycle
>
> Sameer,
>
> Your statement about the analog solver executing at every digital event
> is not correct.
> The execution of the analog solver is related to cycles, not events.
> There are cycles with more than one event, and there are cycles with no
> event at all.
>
> Regarding the second part of your question, I direct you again to
> section 12.6.6 of the LRM, which defines what the execution of the
> analog solver involves. You should be able to draw the right conclusion
> from this section or point to an ambiguity in the definition that
> prevents you from reaching a conclusion.
>
> Ernst Christen
>
> On Thu, 4 Jan 2007 14:51:29 -0500, Sameer Kher wrote:
>> Thanks Ernst,
>>
>> So I am correct in my interpretation of the simulation cycle that the
>> analog solver must be executed at every digital event (perhaps
>> excluding some delta cycles) irrespective of whether the digital event
>
>> affects the analog system or not?
>>
>> Thanks John and Martin for the information about the VHPI.
>>
>> Sameer
>>
>> -----Original Message-----
>> From: Ernst Christen [mailto:Ernst.Christen@synopsys.COM]
>> Sent: Thursday, January 04, 2007 2:02 PM
>> To: Sameer Kher; vhdl-ams@vhdl.org
>> Subject: Re: VHDL-AMS PLI and a Question about the Simulation Cycle
>>
>> Sameer,
>>
>> Step a) defines that "The analog solver is executed." Therefore, the
>> analog solver executes at the beginning of each simulation cycle.
>> Clause
>> 12.6.6 then defines the actions that the analog solver takes when it
>> executes. These actions depend on the result of step
>> g) and other steps, more specifically whether the break flag is set
>> and, to some extent, whether this is a delta cycle or not.
>>
>> Ernst Christen
>>
>> On Thu, 4 Jan 2007 10:23:07 -0500, Sameer Kher wrote:
>>> Also, I was seeking a clarification on the simulation cycle as
>>> described in the LRM. My question is related to step (g) where the
>>> next simulation time is determined. The text says that the next
>>> simulation cycle "...is determined by setting it to the earliest of :
>>> - The value of type universal_time corresponding to TIME'HIGH
>>> - The next time at which a driver becomes active, or
>>> - The next time at which a process resumes..."
>>> Does this imply that the analog solver must execute at every digital
>>> event?
>>>
>>> Thanks,
>>> Sameer Kher
>>> R&D Engineer
>>> Ansoft Corp.


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