From nowick@cs.columbia.edu Wed May 10 23:52:28 1995 Date: Wed, 10 May 95 18:47:15 EDT From: Steven Nowick To: asynchronous-private@cs.columbia.edu Subject: ["Alex Semenov" : A question about timing analysis] Reply-To: asynchronous@hohum.stanford.edu Content-Length: 4322 X-Lines: 95 Status: RO Received: from cs.columbia.edu (cs.columbia.edu [128.59.10.13]) by ober.cs.columbia.edu (8.6.12/8.6.6) with ESMTP id HAA08586 for ; Thu, 4 May 1995 07:43:29 -0400 Received: from cheviot.ncl.ac.uk (cheviot.ncl.ac.uk [128.240.2.10]) by cs.columbia.edu (8.6.12/8.6.6) with ESMTP id HAA17635 for ; Thu, 4 May 1995 07:40:49 -0400 Received: from newton.ncl.ac.uk by cheviot.ncl.ac.uk id (8.6.10/ for ncl.ac.uk) with SMTP; Thu, 4 May 1995 12:40:44 +0100 From: "Alex Semenov" Message-Id: Subject: A question about timing analysis To: nowick@ober.cs.columbia.edu Date: Thu, 4 May 1995 12:40:34 +0100 (BST) Cc: n384758%tanhill@newcastle.ac.uk (me on tanhill) X-Mailer: ELM [version 2.4 PL24] Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII Content-Length: 3363 Dear Steve, I am not quite sure how the asynchronous mailing list works at the moment. Could you please post my message there for me? Thank you in advance. Cheers, Alex. -------------------------------------------------------------------------- EMAIL: Alex.Semenov@newcastle.ac.uk Alex Semenov, PHONE: +44 191 222 6053 Department of Computing Science, FAX : +44 191 222 8232 University of Newcastle upon Tyne, TELEX: uk+53654-UNINEW_G Newcastle upon Tyne, NE1 7RU, UK -------------------------------------------------------------------------- -----------------%----------%----------%----------%----------%------------ Hello everyone, My question is probably addressed to those of you who work with timing analysis using Petri net formalisms, although those who don't are welcome to comment as well. I am studying timing analysis of asynchronous circuits based on Petri net formalisms. It is obvious that there exist two approaches: assign time delays on places and assign delays on transitions. My question arises from the following example: N1 N2 p1 p1 [0,0] /\ /\ / \ / \ / \ / \ [1,5] t1 t2 [2,6] t1 t2 | | | | | | | | | | | | p2 p3 [1,5] p2 p3 [2,6] These two nets model different behaviour: In both models, both transitions are enabled immediately when p1 has a token. However, the firing of t1 and t2 in N1 is delayed by their non-zero firing time, which is with the given ranges. Therefore, the choice in N1 is made with some delay from the time of marking p1 with a token, that is only at the time of firing of (exactly) ONE of the transitions. In N2, the choice is made immediately when p1 has a token, but then the subsequent action is delayed by having a delay in marking p2 or p3. Place p3 in the net N1 will be marked with a maximum delay of 5 (not 6!) after p1, whereas place p3 in the net N2 will be marked with the maximum delay of 6 from p1. Thus, in N1 transitions are assumed to fire within the specified range and in N2 places are assumed to delay appearance of a token in them for the specified range. However, it seems that one can draw a net with timing ranges assigned to transitions modelling behaviour N2 and a PN with timing ranges assigned to places modelling N1 (using additional transitions and places). My question is: How and why do you choose one model over another? Or, if you are using a different model, then which and why? Any other comments are also appreciated. Cheers, Alex. PS I have read works by Alur et al., Rokicki, Nielsen et al., Yoneda et al., Myers et al., Hulgaard et al. Any further references would also be helpful. -------------------------------------------------------------------------- EMAIL: Alex.Semenov@newcastle.ac.uk Alex Semenov, PHONE: +44 191 222 6053 Department of Computing Science, FAX : +44 191 222 8232 University of Newcastle upon Tyne, TELEX: uk+53654-UNINEW_G Newcastle upon Tyne, NE1 7RU, UK --------------------------------------------------------------------------