Re: [sv-ac] Two items for ease of assertion writing

Thomas,
  On #1: this is indeed a trap/nitty gritty, we discussed this in our
books as guideline and also cover this in our regular SVA trainings.
this usually is an issue for assertions involving combinatorial
signals and for simple non-temporal properties (without delays I
mean). We don't see them as many as to warrant a language change, but
that's just one view. We believe this could be a good lint rule for a
SVA linter if one were to create one.

On #2: if I understand you correctly, you may be able to do this with
a delayed version of txn-id signal maybe? Sorry if I misunderstood it.

Regards
Srini
www.cvcblr.com

On 4/13/10, Thomas Thatcher <thomas.thatcher@oracle.com> wrote:
> Hi Dmitry,
>
> I have two question that my turn into enhancement
> requests/clarifications if necessary.
>
> 1. Use of sampled values in action blocks: It appears that to get
> meaningful messages in action blocks, you must explicitly $sample() all
> variables that are used in the message, or you will get interesting results
>
> for example
> assert property (@(posedge clk) req |-> req_id != 4'b0) else
> $display("Error, 0 is an invalid ID: req_id=%d", req_id);
>
> If the sampled value is 0, but the signal changes to 1 in the same time
> step, you'll get the following message:
> Error, 0 is an invalid ID: req_id=1
>
> (I guess one approach would be to write the assertions inside a checker)
>
>
>
> 2. Is there a way to index an array enclosed by $past() using the
> current sampled value of a variable?
>
> for example
> txn_valid(txn_id) && $past(txn_valid[txn_id])
> ^want to index with the
> current sampled value
> of txn_id
>
> The only way I have figured out to make this work is to explicitly
> create a delayed version of txn_valid and use that instead of the
> $past() function.
>
> Another option I thought of just now would be to extend the $stable()
> system function to take an optional parameter for number of cycles.
>
>
> I can enter Mantis items if either of these turns out to be more than
> just a user education issue.
>
> Thanks,
>
> Tom
>
>
> On 04/12/10 10:15, Korchemny, Dmitry wrote:
>> Hi all,
>>
>>
>>
>> In the reminder I sent was written that the meeting would take place in
>> two weeks. This is a copy-paste mistake. The meeting will take place
>> *tomorrow*. The date in the invitation Date: 2010-04-13 Time: 16:00 UTC
>> (09:00 PDT) is correct.
>>
>>
>>
>> Thanks,
>>
>> Dmitry
>>
>> ---------------------------------------------------------------------
>> Intel Israel (74) Limited
>>
>> This e-mail and any attachments may contain confidential material for
>> the sole use of the intended recipient(s). Any review or distribution
>> by others is strictly prohibited. If you are not the intended
>> recipient, please contact the sender and delete all copies.
>>
>>
>> --
>> This message has been scanned for viruses and
>> dangerous content by *MailScanner* <http://www.mailscanner.info/>, and is
>> believed to be clean.
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
>

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