Re: [sv-ac] sequence as clocking event to covergroup

Hi Tej:

I have a few comments.

It is true that the LRM does not give a complete definition of
multiplicity of matching and of when matching occurs.  On the other
hand, the LRM also does not leave these things completely undefined.

I think that if a covergroup has a sequence instance as its clocking
event, then multiple matches of the sequence instance in the same
timestep should result in multiple triggerings of the covergroup.

I think we should avoid allowing the rules for multiplicity of covergroup
triggering on sequences to differ from the rules for multiplicity of
matching that apply in other scenarios, e.g.

   cover property (
      @(posedge clk)
      sequence_instance ##0 (1, task_call);
   );

I don't like the idea of trying to incorporate first_match into the
semantics of covergroups with sequence clocking event.  I think the 
user can use first_match in defining a sequence, but then the general
rules for multiplicity of sequence matching should apply.

In particular, a first_match can have multiplicity, e.g. if both
a ##1 b and c ##1 d match over a two-cycle interval, then 

   first_match((a ##1 b) or (c ##1 d))

should match with multiplicity two over that interval.

Best regards,

John H.

> X-Authentication-Warning: server.eda-stds.org: majordom set sender to owner-sv-ac@eda.org using -f
> X-MimeOLE: Produced By Microsoft Exchange V6.5
> Content-class: urn:content-classes:message
> Date: Wed, 6 Dec 2006 15:12:02 -0800
> X-MS-Has-Attach: 
> X-MS-TNEF-Correlator: 
> Thread-Topic: [sv-ac] sequence as clocking event to covergroup
> Thread-Index: AccZfiKOmZ5lYIN/QgGxOLcn04V72QABWXHgAAFLp5AAAHSA7A==
> From: "Singh, Tej" <tej_singh@mentor.com>
> Cc: <sv-ac@eda.org>
> X-OriginalArrivalTime: 06 Dec 2006 23:12:03.0364 (UTC) FILETIME=[F0C06E40:01C7198B]
> X-Virus-Status: Clean
> Sender: owner-sv-ac@eda.org
> 
> This is a multi-part message in MIME format.
> 
> ------_=_NextPart_001_01C7198B.F0397F78
> Content-Type: text/plain;
> 	charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
> 
> But as Doron pointed out, the LRM does not define the number of threads.
> 
> How about defining it so that the number of events at a clock is same as
> number of sequence attempts that matched at that clock. This is what I =
> was trying
> to do with first_match but first_match will kill any future matches.
> 
> So for a sequence expr that looks like
> 
> a ##[1:$](b ##1 c) or (d ##1 e)
> 
> it will not matter whether 'or' results in two matches or one match =
> since
> both correspond to the same attempt.=20
> 
> Tej
> 
> 

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