( ESNUG 467 Item 4 ) -------------------------------------------- [07/26/07]
Subject: ( ESNUG 465 #3 ) Why is the DC-to-Formality guidance a "secret"?
> This user is trying to conclude guidance files of any kind are always a
> bad idea -- but it seems that his experiment did NOT include comparing
> "with guidance" to "without guidance." That's the essential comparison,
> and this poster hasn't even tried it in the tool he uses (ConFormal),
> let alone the tool he is drawing a conclusion about Formality.
>
> Any Formality user can try the experiment: Remove a flop from your netlist
> as if the Design Compiler had mistakenly removed it. Then hand-edit the
> SVF to wrongly instruct Formality that this flop was removed because it's
> a constant register. When I tried it, Formality rejected the wrong
> guidance, said it was doing so, then went on to fail the netlist as
> non-equivalent.
>
> - [ The Silver Surfer ]
From: Tom Mannos <tjmanno=user domain=sandia.gov>
Dear Silver Surfer,
I would like to try the experiment in question. However, when I examine the
SVF file, I see that is a binary file in some proprietary format that I
don't have access to. This is my one and only objection to Formality's use
of guidance files. Knowing that Formality ignores bad guidance isn't as
important to me as knowing that DC doesn't generate bad guidance in the
first place. Again, I don't mind the tools passing notes, as long as
they're not secret notes!
Is there a flag to make DC write out a text SVF? If so, I'd be happy to use
it. Otherwise, I'm stuck doing things the hard way.
- Tom Mannos
Sandia National Labs. Albuquerque, NM
Index
Next->Item
|
|
