Re: [sv-ac] Re: nonblocking assignment in clocking block.

Subject: Re: [sv-ac] Re: nonblocking assignment in clocking block.
From: Adam Krolnik (krolnik@lsil.com)
Date: Fri Sep 19 2003 - 14:11:07 PDT

Hi Surrendra;

You wrote:
>> 3. Use of always blocks is better to model additional logic.

>You can do this as long as you don't need temporal functions and ended.

Now I'm confused.

Arturo is proposing the ability to test a property/sequence for success
(and failure?) - Isn't the ended method call the same thing?

So, if one can write

wait (aproperty.triggered) // Arturo's current proposal

Then why can't someone write:

always @(posedge clk)
   begin
   if (aproperty.ended)
     inprogress <= 1'b1;
   ...

Also, why in a procedural block with a defined clock (by the rules for clock inferring)
could you not use the $past or $sampled functions?

Any variables set in a clocking domain are executed in the reactive region,
after the properties have been evaluated. It would seem that one could
use always blocks/code just as effectively.

I still think only supporting a nonblocking assignment is too minimalistic of a
proposal at a language design level. It would be a fine restriction in a
quarterly released tool though :)

Arturo:

I would be interested in refining the EXT-7 Proposal to define the difference between
    .ended and .triggered and possible add a .failed and possibly a method to
    tell how many threads of a property are executing (or if any are executing.)

    Thanks Surrendra, Arturo;

    Adam Krolnik
    Verification Mgr.
    LSI Logic Corp.
    Co-author "Assertion Based Design"

This archive was generated by hypermail 2b28 : Fri Sep 19 2003 - 14:12:00 PDT