Re: [tlmwg] Revisit of the TLM2.0 phases rules

I think we agree, but you seem to equate timing annotation with temporal decoupling? To me, the decoupling is a property of the simulation scheduling. The length of the time slice depends on what you want to observe. Timing annotation is the sane way to do models, as it makes them much more efficient and separates out timing into something explicit. I was only referring to how you run the simulation not how nodels are written.

/jakob

Jakob Engblom, PhD. Technical Marketing Manager - Simics
Wind River, Stockholm, Sweden
Mobile +46 734 368 958
(Sent from my Blackberry)

----- Original Message -----
From: Aldis, James [mailto:j-aldis2@ti.com]
Sent: Thursday, January 06, 2011 05:21 AM
To: Engblom, Jakob; Robert Guenzel <robert.guenzel@greensocs.com>; Veller, Yossi <Yossi_Veller@mentor.com>
Cc: tlmwg@lists.systemc.org <tlmwg@lists.systemc.org>; systemc-p1666-technical@eda.org <systemc-p1666-technical@eda.org>
Subject: RE: [tlmwg] Revisit of the TLM2.0 phases rules

>
Texas Instruments France SA, 821 Avenue Jack Kilby, 06270 Villeneuve Loubet. 036 420 040 R.C.S Antibes. Capital de EUR 753.920

-----Original Message-----

> From: tlmwg@lists.systemc.org
> [mailto:tlmwg@lists.systemc.org] On Behalf Of Engblom, Jakob
> Sent: Wednesday, January 05, 2011 3:41 PM
> To: Robert Guenzel; Veller, Yossi
> Cc: tlmwg@lists.systemc.org; systemc-p1666-technical@eda.org
> Subject: RE: [tlmwg] Revisit of the TLM2.0 phases rules
>
> > And since we have temporal decoupling at AT, in-order with
> respect to
> > timing would mean to have PEQs everywhere, and that would
> render the
> > whole idea of temporal decoupling meaningless (but I confess that I
> > always wondered if temporal decoupling at AT is really useful...).
>
> In my experience, you cannot really combine temporal
> decoupling to any useful extent with trying to model a shared
> blocking interconnect. The biggest interval you can use is
> something smaller than the minimal blocking time, and even
> anything bigger than one might cause unfairness in which
> requestors get access to the resource.
>
> /jakob
>

I am strong believer that one should try to make one's models as
integration-neutral as possible. I agree that timing annotation may
serve little purpose in a system simulation that is entirely AT,
because it gets removed by PEQs almost immediately. But any
component model might be pulled out of that environment and dropped
into a more loosely-timed simulation. So if the component model
is sufficently inaccurate that delays can be calculated, I
recommend calculating them, using timing annotation, and letting the
other guy decide whether a wait() is needed.

James

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