[sv-ac] RE: 1698

Hi John,

 

I reviewed the 1698. This is completely different than what I thought
was desired for sampled value operation:

 

The value of the expression sampled in the Preponed region of the
current timestep is compared to the value of the expression sampled in
the Preponed region of the most recent strictly prior timestep in which
the clocking event occurred. Here, the current timestep refers to the
simulation timestep

in which the function is called.

 

As another example,

     always_ff @(posedge clk1)

          reg1 <= $rose(b,@(posedge clk2));

 

Here, reg1 is updated in each timestep in which posedge clk1 occurs,
using the value returned from the $rose sampled value function in that
timestep. $rose compares the sampled value of the LSB of b from the
current timestep (one in which posedge clk1 occurs) with the sampled
value of the LSB of b in the strictly prior timestep in which posedge
clk2 occurs.

 

I had pictured $rose(b,@(posedge clk2)) as comparing two values, both of
which occurred at posedge clk2. The way you have written it, we compare
the current value at posedge clk1 with the previous sampled value when
clk2 occurred.

 

This was not my understanding of what was desired though it is another
possible implementation.

 

As an aside, I have noticed that LSB (and lsb) refers to least
significant bit in most places of the LRM but on page 207 it is least
significant byte. That has caused some confusion, though it is obvious
to me. Should we specify it?

 

Lisa

 

-----Original Message-----
From: John Havlicek [mailto:john.havlicek@freescale.com] 
Sent: Monday, February 11, 2008 8:47 AM
To: sv-ac@eda.org
Cc: doron.bustan@intel.com; eduard.cerny@synopsys.com;
yaniv.fais@freescale.com; john.havlicek@freescale.com;
dmitry.korchemny@intel.com; Manisha_Kulshrestha@mentor.com;
johan.martensson@jasper-da.com; Lisa Piper; erik.seligman@intel.com;
bassam.tabbara@synopsys.com; thomas.thatcher@sun.com
Subject: 1698

 

Hi Folks:

 

On Friday I posted a revision of 1698 that I think is

in good shape.

 

I believe that Ed's characterization of the RTL was correct,

but it is not necessary to talk about it in order to define

what the sampled value functions mean.  Since it was raising

concern, it has been deleted.

 

I plan to call for e-mail vote on this for our next review 

round.

 

J.H.


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