Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!simulation From: simulation@uflorida.cis.ufl.edu (Moderator: Paul Fishwick) Newsgroups: comp.simulation Subject: SIMULATION DIGEST V13 N3 Message-ID: <21607@uflorida.cis.ufl.EDU> Date: 3 Jan 90 14:46:33 GMT Sender: fishwick@uflorida.cis.ufl.EDU Reply-To: simulation@uflorida.cis.ufl.edu Lines: 241 Approved: fishwick@uflorida.cis.ufl.edu Volume: 13, Issue: 3, Wed Jan 3 09:46:20 EST 1990 +----------------+ | TODAY'S TOPICS | +----------------+ (1) UDL/I High Level Design Language (2) TR: Determining GVT in Time Warp Simulation (3) TR: Protocols for Parallel Logic-Level Simulation (4) Simulation of Data Flow Architectures (5) Simulation for Factory Automation * Moderator: Paul Fishwick, Univ. of Florida * Send topical mail to: simulation@bikini.cis.ufl.edu OR post to comp.simulation via USENET * Archives available via FTP to bikini.cis.ufl.edu, login as 'ftp', use your last name as the password, change directory to pub/simdigest. * Simulation Tools available by doing above and changing the directory to pub/simdigest/tools. ----------------------------------------------------------------------------- Posted-Date: Tue, 19 Dec 89 18:59:55 CST Date: Tue, 19 Dec 89 18:59:55 CST From: steve@titan.tsd.arlut.utexas.edu (Steve Glicker) To: simulation@ufl.edu Subject: The UDL/I HDL In the December 18th issue of EE Times a front page article entitled "Japan Plans HDL" begins, "Tokyo - A powerful group of Japanese companies, spearheaded by NTT Corp., is hammering out a high-level design language called UDL/I. When complete, the Unified Design Language for ICs will be proposed as a Japan Industrial Standard, raising questions about the future of VHDL, the emerging hardware-description language standard." ... The article indicates that UDL/I is in its early stages. Does anyone know how to obtain a draft of this proposed standard? Steve Glicker (steve@titan.tsd.arlut.utexas.edu) ------------------------------ Date: Tue, 26 Dec 89 11:49:03 -0800 From: liny@cs.washington.edu (Yi-Bing Lin) Return-Path: To: fishwick@bikini.cis.ufl.edu Subject: Determining GVT in Time Warp simulation The following technical report (TR 90-01-02) can be requested via e-mail: tech-report@june.cs.washington.edu DETERMINING THE GLOBAL VIRTUAL TIME IN A DISTRIBUTED SIMULATION Yi-Bing Lin and Edward D. Lazowska Department of Computer Science and Engineering University of Washington Seattle, WA 98195 ABSTRACT The virtual time paradigm is a method of organizing and synchronizing distributed systems. An implementation of this paradigm, called the Time Warp mechanism, is one of the most important parallel simulation protocols. The Time Warp synchronization mechanism takes an optimistic approach in which a process executes every message as soon as it arrives. If a message with an earlier timestamp subsequently arrives, the process rolls back its state to the time of the earlier message and re-executes from that point. Thus, the state of each process must be saved regularly (regardless of whether or not rollbacks actually occur). The amount of storage used for state-saving grows as the simulation progresses. Jefferson observed that at any real time there exists a global virtual time GVT such that all executed messages with timestamp earlier than GVT will not be rolled back. Thus the storage used for saving information with timestamp earlier than GVT can be reclaimed. In addition to garbage collection, GVT can be useful in other areas of Time Warp simulation such as termination detection, snapshots and crash recovery, input and output handling, etc. The task of finding GVT is not trivial in a distributed environment. The approach used in most distributed Time Warp implementations is based on Samadi's algorithm. This algorithm requires acknowledgement messages, which introduce heavy network traffic. This paper proposes a new algorithm that does not require acknowledgement messages, eliminating 50\% of the network traffic in Time Warp simulation. A data structure used in our algorithm can also be used to detect lost messages in an unreliable environment. ------------------------------ Date: Tue, 26 Dec 89 11:56:49 -0800 From: liny@cs.washington.edu (Yi-Bing Lin) Return-Path: To: fishwick@bikini.cis.ufl.edu Subject: protocols for parallel logic-level simulation The following technical report (TR 89-09-06) can be requested via e-mail: tech-report@june.cs.washington.edu COMPARING SYNCHRONIZATION PROTOCOLS FOR PARALLEL LOGIC-LEVEL SIMULATION Yi-Bing Lin and Edward D. Lazowska Department of Computer Science and Engineering University of Washington Seattle, WA 98195 Mary L. Bailey Department of Computer Science University of Arizona Tucson, AZ 85721 ABSTRACT Recently there has been a great deal of interest in parallel event-driven logic-level simulation as a means of alleviating the simulation bottleneck in VLSI design. This paper compares different synchronization protocols for parallel logic-level simulation. The protocols investigated are the synchronous protocol, the Chandy-Misra protocol, the Time Warp protocol with aggressive cancellation, and the Time Warp protocol with lazy cancellation. An artificial protocol, called the {\it conservative optimal protocol}, is used as a basis for this comparison. Under the assumption that every process is assigned to a dedicated processor, we show that the conservative optimal simulation outperforms both the synchronous simulation and the Chandy-Misra simulation, and that the Time Warp simulation with aggressive cancellation and the Time Warp simulation with lazy cancellation both outperform the conservative optimal simulation. ------------------------------ To: comp-simulation@rutgers.edu Path: bimacs!ariel From: ariel@bimacs.biu.ac.il (Ariel J. Frank) Newsgroups: comp.simulation Subject: Simulation of (data flow) architectures Keywords: simulation, architecture, dataflow, software Date: 21 Dec 89 09:36:33 GMT Organization: Math & CS, BarIlan U, Ramat-Gan, Israel Hi Simulation Land. I'm new to this group so please indilge with me. I need your collective advice on a suitable PD (or reasonablly priced) software system for developing a simulation program and/or simulating computer architectures and especially data flow oriented ones. The idea is to simulate at the object level but objects can be at a low level and there can be many of them. The setup has to be well documented and easily usable (for architecture guys). Preferences to UNIX setups for VAXen, VAXStations or SUNs. Any info, facts, advice, insight will be appreciated. Please E-mail and I will summerize if enough interest. Thanks in advance, Ariel. -- Ariel J. Frank Deputy Chairperson, Dept. of Mathematics and Computer Science Bar Ilan University, Ramat Gan, Israel 52100 Tel: (972-3-) 5318407/8, Fax: (972-3-) 344766 BITNET: ariel@bimacs (also F68388@barilan) INTERNET: ariel@bimacs.biu.ac.il ARPA: ariel%bimacs.bitnet@cunyvm.cuny.edu CSNET: ariel%bimacs.bitnet%cunyvm.cuny.edu@csnet-relay UUCP: ...uunet!mcvax!humus!bimacs!ariel ------------------------------ To: uunet!comp-simulation@uunet.UU.NET Path: cme!durer!newton From: newton@cme.nist.gov (Eric Newton) Newsgroups: comp.simulation Subject: novice questions on available software Date: 2 Jan 90 18:26:07 GMT Sender: news@cme.nist.gov Distribution: comp Organization: National Institute of Standards and Technology Hi, I am hoping that some of you experts out there in netland could help me. A group here is attempting to build a system to integrate software packages for factory automation. Simulation is one of the systems that we would like to integrate. Unfortunately, none of us has much experience with simulation packages currently available. I am soliciting your advice for such packages. We are *not* looking for 3D graphics packages that will simulate machinery or robots moving on the shop floor. We *are* looking for packages that will simulate the data flow and timing requirements of a system. We have a variety of resources to run such a system on: Macs, PCs, Sun 3s, SGI Personal 4D, a Symbolics lisp machine. The cost of the software is important, but the flexibility of the system should have highest priority. Public Domain(ish) code or source is highly desirable. We could use any advice/opinions you might have. -Eric -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ newton@cme.nist.gov Eric C. Newton Computer Scientist (Looking for something witty to put here) NIST (formerly NBS) Gaithersburg, MD 20899 ------------------------------ END OF SIMULATION DIGEST ************************