RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

From: Alsop, Thomas R <thomas.r.alsop_at_.....>
Date: Tue Jan 29 2008 - 13:46:26 PST
Hi Doug,

 

Sorry it's taking me so long to get around to implementing this. Two
week long business trip training teams all over the place combined with
jet lag and I'm a bit frazzled right now.  I will try and get this done
in the next couple days so we can talk about it during the next
conference call.  

 

I haven't read the 2005 proposal yet.  I'll take a look at it tomorrow.
I'm hoping to glean from it the wording they used to describe the "false
glitch" avoidance.  Eric explained it to me so I understand the
semantics, just want to make sure my wording is right.

 

I met with Shalom on Sunday in Israel.  He explained the assertion issue
to me.  I personally don't like this falling into the category of
assertions because I don't want it turned off, at least I don't see a
reason to disable it.  If there is a strong argument on both sides
however, then we can leave it vague.  I have no issue with that.

 

Thanks for taking these notes for me and thanks for the Rev5 update.
Rev6 coming soon.

 

-Tom

 

 

________________________________

From: Warmke, Doug [mailto:doug_warmke@mentor.com] 
Sent: Monday, January 21, 2008 5:41 PM
To: Alsop, Thomas R; sv-bc@server.eda.org
Subject: RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

 

Hi Tom,

 

In the SV-BC today we discussed 2008 at some length.

The consensus was that the proposal needs to define more precise

semantics of how the avoidance of "false glitches" is to be implemented.

 

Another topic was "Are unique/priority if/case considered to be SV
assertions of some kind?"

Matt M would like that to be the case.  Stu and others argued it should
not be the case.

In the end, I think the idea was to leave that matter explicitly vague.
That way tools could

have leeway to provide command-line switches that swing the constructs
one way or the

other (i.e. optionally allow $assertoff and $assertkill to affect
unique/priority violation checks)

Others on the committee surely will have something to add here.
Personally I have no strong

opinion; I'm just trying to relay what I heard at the meeting.

 

I've uploaded a rev5 of the 2008 proposal which takes care of a lot of
the above.

The main idea is to borrow relevant sections of 2005's proposal and
mechanics.

If 2005 changes substantially (which I doubt at this point), then 2008
would need

to be updated with similar changes.

 

The rev5 proposal is not complete.  I mainly wanted to give a running
start to things

based on my experience with 2005 and the LRM at large. Some example work
still

needs to be done in 2008 (presumably by you).  Also, I only made the
changes in the

unique/priority if section.  The same kinds of changes would need to

be done in the unique/priority case section.

 

Thanks,

Doug

 

 

From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org] On
Behalf Of Alsop, Thomas R
Sent: Sunday, January 20, 2008 8:19 PM
To: Alsop, Thomas R; Brad Pierce; sv-bc@server.eda.org
Subject: RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

 

I have updated the proposal for 2008 (see Rev4) to remove basically all
references to 2005, removing the #0 and throughput qualifiers.  This has
greatly simplified the proposal.  I am not sure if I'll log on tonight
for the meeting mostly because it will be very late at night here in
Bangalore _and_ I am jet lagged.  Matt or Shalom, if you can make notes
about any comments from the latest proposal, I can address them over the
next week and we can vote again on this sometime ww05.  

 

Thanks, -Tom

 

________________________________

From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org] On
Behalf Of Alsop, Thomas R
Sent: Tuesday, January 15, 2008 11:10 PM
To: Brad Pierce; sv-bc@server.eda.org
Subject: RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

 

I agree with Brad on issue 5 below.  Can we just remove "throughout"
from this proposal?  Does anyone want that in there? I don't care about
it.  If you do, is there a recommendation for a better keyword? Thanks,
-Tom

 

________________________________

From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org] On
Behalf Of Brad Pierce
Sent: Tuesday, January 15, 2008 7:01 AM
To: sv-bc@server.eda.org
Subject: RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

 

Doug,

 

I agree with your objections 1-3 about the current version.  I'm happy
though with the compatibility break you mention in 4 (from default
'throughout' to default '#0'), because unique/priority themselves are
already broken.  A precedent is the backward incompatible fixes to
'generate' in 2005.

 

    5)  I also object to the re-use of the 'throughout' operator as a
qualifier entirely unrelated to its conventional meaning in SVA
sequences.  I suppose the intent is 'throughout the time step'.  But I
doubt anyone will really want this 2005 immediate-style semantics going
forward anyway.

 

I recommend not changing the syntax of unique/0 and priority, but only
changing the semantics.

 

-- Brad

 

________________________________

From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On Behalf Of
Warmke, Doug
Sent: Tuesday, January 15, 2008 12:22 AM
To: Maidment, Matthew R; sv-bc@eda.org
Subject: RE: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

Hello all,

 

I vote No on Mantis 2008 for the following reasons:

 

1)      The related Mantis 2005 has been rewritten such that the
event-control
form of the syntax is no longer present.  That had too many unresolved
issues,

so SV-AC decided to postpone that enhancement until sometime in the
future.

Thus, the event-control aspect of 2008 should be removed, too.

2)      The example in the unique/priority if area specifically mentions
a 4 ns delay.
But that is not actually the case in the example.  Rather, this is an
example
that is immune to zero-delay glitches in the active region set.  Note
that 
evaluation of the unique-ness/priority-ness of the conditions is
supposed

to happen in the Observed region, as per alignment with the deferred
assertion feature of Mantis 2005.  Thus, "zero-delay glitch" isn't quite
an

accurate term.  It should be "zero-delay glitch in the active region
set".

(Since oddball glitches caused by zero-delay oscillations across the

active and reactive region sets would still fire the violation checks)

3)      Speaking of "violation checks", I would prefer it if 2008 caused
that wording

to be used when describing unique/priority if/case.

4)      I'm not in favor of the compatibility break.  I think that the
proposed default
behavior is too sophisticated to be allowed without the #0 syntactic
clue.
It's not hard to add those #0 into the source code, and it does give the
reader the clue that unique/priority violations will be checked with
some
zero-delay semantic.  In addition, the current version of the construct
can

work fine if placed in clocked procedures that include logic when
assigning
the clocked output variables of the procedure.  (Thus, the current
semantics

aren't totally useless, though I do agree they are pretty useless for
combinational

procedures)

 

I'd like to hear what others have to say about 4).

If there was enough weight in favor of making the compatibility break,

I will lift this particular objection, since I do think 2008 has a lot
of value

and should be passed in this version of the standard.

 

Regards,

Doug

 

From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org] On
Behalf Of Maidment, Matthew R
Sent: Monday, January 14, 2008 4:53 PM
To: sv-bc@server.eda.org
Subject: [sv-bc] E-mail ballot: DUE 8am PST, Jan 21, 2008

 

 

-You have until 8am PST, Monday, January 21, 2008 to respond 
-An issue passes if there are zero NO votes and half of the eligible 
 voters respond with a YES vote. 
-If you vote NO on any issue, your vote must be accompanied by a reason.

 The issue will then be up for discussion during a future conference
call. 
-Note: For some issues, the proposed action is captured in the bug note 
       (resolve as duplicate, already addressed, etc.). 

As of the January 7, 2008 meeting, the eligible voters are: 

Brad Pierce        
Shalom Bresticker  
Cliff Cummings      
Mark Hartoog        
Francoise Martinolle 
Karen Pieper       
Dave Rich          
Steven Sharp       
Gordon Vreugdenhil 
Stu Sutherland 
Alex Gran 
Don Mills 
Heath Chambers 
Tom Alsop 
Doug Warmke 
Mike Burns 

SVDB 2008 ___Yes   ___No 
http://www.eda.org/svdb/view.php?id=2008
<http://www.eda.org/svdb/view.php?id=2008>  

 


-- 
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 <http://www.mailscanner.info/> , and is

believed to be clean. 


-- 
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 <http://www.mailscanner.info/> , and is

believed to be clean. 
-- 
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.
Received on Tue Jan 29 13:49:10 2008

This archive was generated by hypermail 2.1.8 : Tue Jan 29 2008 - 13:49:54 PST