MONTHLY PROGRESS REPORT CONTRACTOR: University of California at Berkeley AGREEMENT NUMBER: DAAB07-97-C-J007 CONTRACT PERIOD: 11/18/96 - 11/31/99 DATE: Feb. 23, 1999 TITLE: Heterogeneous Modeling And Design REPORT PERIOD: 12/15/98 - 1/15/99 SPONSOR: Air Force Research Laboratory (AFRL) TECHNICAL POC: James P. Hanna REPORT PREPARED BY: Edward A. Lee 0. Executive Summary Technology transfer continues unabated, with a new product based on Ptolemy software by Virtual Photonics, and reports of use of our software in products by LavaLogic and Thomson CSF. We also filed four masters reports in this period, covering continuous-time modeling, communicating sequential processes with a timed extension, process-network modeling, and protocol modeling. All four reports are available on the Ptolemy web site. We have also scheduled a Ptolemy Miniconference to be held at Berkeley on Friday, February 19. 1. Research Status Task 1: Modular deployable design tools ======================================= Plot package ------------ The plot package has been upgraded to use double buffering for graphics display so that animated plots that are built by doing "clear()" between data sets are displayed smoothly, rather than flashing. Also, it now has an "export" facility that produces an encapsulated postscript rendition of the plot. Other Software -------------- Christopher Hylands fixed a number of bugs in Tycho that date back to the upgrade to Itcl 3.0. He also made a number of changed to Ptolemy 0.7.2 so that it builds under Cygwin b20.1 and under Microsoft Visual C++ 6.0. This shows that Ptolemy 0.x is portable and it can be used (with some effort) on a non-Unix platform. He also began experimenting with using InstallWizard to wrap Ptolemy software distributions, but ran into difficulties probably due to the very large size of Ptolemy 0.x. He experimented with using JShield, InstallWizard's Java installation tool to build a self extracting Java installer for ptplot. This effort should lead to a very easily installed Ptolemy II. Task 2: Domain-specific design tools ==================================== Math Package ------------ Jeff Tsay has dramatically improved the speed of the FFT implemented in the SignalProcessing class of the math package. He has also constructed thorough regression tests for it. Software Practice ----------------- Design reviews: - ct kernel package - pn kernel package Task 3: Heterogeneous interaction semantics =========================================== SDL --- Our regular Friday studygroup is studying SDL (Specification and Description Language), a popular modeling language in telecommunications. Multi-Tool Modeling ------------------- William Wu and Jie Liu have implemented a Java interface to the Saber analog simulation tool. Jie is in the process of making a CT Actor with it so that Saber simulations can be included as components in the CT (continuous-time) domain. 2. Equipment/Infrastructure Status: We have ordered four new Sun UltraSparc machines. 3. Interactions and Technology Transfer -- Harald Hamster of Virtual Photonics sent us an announcement of the Photonics Transmission Design Suite - PTDS (http://www.virtualphotonics.com/products/ptds/ptds.html), which is based on Ptolemy 0.7. He provided the following background: "In the early 1990s, device performance constrained system capabilities. Beam Propagation Modeling emerged to advance integrated optics performance. By 1996, transmission speed had grown to the point where non linearities in the fiber channel emerged as the critical performance factor. A second generation of tools was developed with symbolic editors and sophisticated fiber models. Now, the aggressive use of fiber bandwidth in systems with 80 WDM channels (and more) is creating new hurdles for design tools, while the first optical metro networks with OADMs and OXCs bring further challenges in system and subsystem complexity. Virtual Photonics Incorporated (formerly BNeD and Virtual Photonics) was created to provide the vision, focus, expertise and resources needed to create the third generation of design tools. Since March 1998, we have invested the best efforts of a twenty strong, design team of simulation and photonics experts to ramp the PTDS to market. This development work was driven by five pressing needs: 1. Huge bandwidth, which PTDS manages with effective new signal approximations 2. Complex ring and interconnected ring networks which PTDS manages with sophisticated schedulers 3. Complex components and subsystems such as OADMs, OXCs and EDFAs, which PTDS manages with hierarchical design capability 4. Codesign of components and systems, which PTDS handles by mixing simulation paradigms 5. Use of in house tools and models, often built with standard tools like Matlab, which PTDS facilitates with a special interface and general data formats via TCP/IP. Since November 1998, a beta test of the LINUX and SUN Solaris versions of PTDS has been underway in 40 companies. During the test period, we have focused development on a new GUI and speed. And we have established a training programs at the Universities of Southern California and Aston to make sure PTDS users get an early and competitive return on the design tool investment." -- Thomson CSF has contributed C code generation domains with both SR (synchronous/reactive) and FSM (finite-state machine) domains. -- Toby Bennett, vice-president and general manager of LavaLogic, of Columbia, Maryland, informs us that they are using Diva in their product. -- Andrew Wendelborn of the University of Adelaide visited our group and gave a talk entitled "Process networks as a high-level notation for meta-computing." -- Paul Caspi, from IRISA, Grenoble, France, visited our group and discussed synchronous languages, clock calculi, and dataflow modeling. 4. Personnel Status John Reekie is leaving the group to join the new GSRC, Gigascale Silicon Research Center, at Berkeley. Will will continue to work with him intensively, but his charter is now broader. 5. Talks/Presentations/Publications: Publications ------------ [1] Jie Liu, Bicheng Wu, Xiaojun Liu, and Edward A. Lee, "Interoperation of Heterogeneous CAD Tools in Ptolemy II," Accepted to Symposium on Design, Test, and Microfabrication of MEMS/MOEMs, March 1999, Paris, France. [2] Edward A. Lee, "Modeling Concurrent Real-time Processes Using Discrete Events," accepted for publication in the Annals of Software Engineering special volume on Real-Time Software Engineering. [3] Neil Smyth, 'Communicating Sequential ProcessesDomain in Ptolemy II," MS Report, UCB/ERL Memorandum M98/70, Dept. of EECS, University of California, Berkeley, CA 94720, December 1998. [4] Jie Liu, "Continuous Time and Mixed-Signal Simulation in Ptolemy II," MS Report, UCB/ERL Memorandum M98/74, Dept. of EECS, University of California, Berkeley, CA 94720, December 1998. [5] Shang-Pin Chang, "System-Level Modeling and Evaluation of Network Protocols," MS Report, UCB/ERL Memorandum M98/73, University of California, Berkeley, CA 94710, December 16, 1998. [6] M. Goel, "Process Networks in Ptolemy II" MS Report, ERL Technical Report UCB/ERL No. M98/69, University of California, Berkeley, CA 94720, December 16, 1998. 6. Difficulties/Problems None to report. 7. Next Quarter Plans We are preparing for the Ptolemy Miniconference on February 19, where we plan to give the first public demonstrations of Ptolemy II. This preparation will involve consolidation of various software packages and conversion of many test cases into demonstratable applets. 8. Financial Data Provided separately on a quarterly basis by the university.