Subject: Re: [sv-ac] Alternative proposal for NBA assignment
From: Adam Krolnik (krolnik@lsil.com)
Date: Fri Nov 14 2003 - 08:05:07 PST
Hi Surrendra;
Reading the sections I have some comments.
In section 1.1.1, I think it would be good to have a waveform picture showing
the values of reg1, a, b and clk. This would ensure that everyone understands
the functionality. One may even want to include a waveform for $rose(b).
In section 1.1.3, you write,
"s1.triggered is true in the simulation time unit after the observe region in
which the sequence match occurs. S1.triggered gets reset to false in the
next simulation time unit."
The phrase "simulation time unit after the observe region" is referring to the
reactive region, correct?
This above sentence and restated as the third bullet in the next paragraph means
that the method "triggered" only works for the reactive region.
I do not understand why you are only considering program block code in trying to
fulfil this functionality need. I do not accept as reasonable the code fragment
to switch to the reactive region.
[Actually, if this is acceptable by everyone else, then we could rewrite the action
block to execute in the active region and use this trick to switch to the reactive
region when one desires.]
I propose that we extend the value of the method "triggered" to be active for one
full clocking cycle - observed .. next clock observed region. This allows one
to write verilog code executing in the active region to be utilized for modeling
code.
What would happen if I wrote:
forever
L3: @(posedge clk) sig3 <= sig4 & s.ended;
We should explain this as it is not defined anywhere what would occur.
Thanks.
Adam Krolnik
Verification Mgr.
LSI Logic Corp.
This archive was generated by hypermail 2b28 : Fri Nov 14 2003 - 08:06:43 PST