2nd ANNUAL RASSP CONFERENCE PAPERS

  1. Advanced Technology Laboratories’ Path to 4X Improvements
  2. Approximate Processing and Incremental Refinement Concepts
  3. Architectures for a RASSP Signal Processor
  4. Autocoding in the Lockheed Martin ATL/Camden RASSP Hardware/Software Codesign
  5. Automated Generation of Accurate VLSI Behavioral Processor Models for Simulation and Synthesis
  6. Collaborative VHDL Modeling within the RASSP Program Demonstration Project
  7. Creating a "Plug and Play" Architecture to Support Hierarchical Virtual Enterprise Design Process Management
  8. Developing Re-useable Performance Models for Rapid Evaluation of Computer Architectures Running DSP Algorithms
  9. Environment and Tools for an Integrated RDE (ENTIRE)
  10. Evaluating Distributed Multiprocessor Designs
  11. A Framework for the Development of Hybrid Models
  12. Hardware/Software Codesign in the Lockheed Martin Advanced Technology Laboratories’ RASSP Program
  13. Implementation of the RASSP SAR Benchmark on the Intel Paragon
  14. Integration of DFT into RASSP
  15. Integrated Process Control and Data Management in RASSP Enterprise
  16. Integrated Reliability and Performance Evaluation using Information Flow Models
  17. KINDLING: A RASSP Application Case Study
  18. Lockheed Martin Advanced Technology Laboratories RASSP Second Year Overview
  19. Managing the RASSP Virtual Enterprise
  20. Projecting RASSP Benefits
  21. Rapid Design and Exploration of Signal Processing System Using a VHDL Model Generator Based Paradigm
  22. Rapid Prototyping of Application-Specific Signal Processors: Educator/Facilitator Current Practice (1993) Model and Challenges
  23. Rapid Prototyping Applied to Underwater Acoustic Modem Research & Development
  24. RASSP Benchmark 1: Virtual Prototyping of a Synthetic Aperture Radar Processor
  25. RASSP Benchmark-1 and -2: A Preliminary Assessment*
  26. RASSP Process Management
  27. RASSP Program Overview
  28. RASSP: The Road to 4X
  29. RASSP Technology Insertion into the Synthetic Aperture Radar Image Processor Application
  30. RASSP VHDL Modeling Terminology and Taxonomy
  31. Real Time IRST Development Using RASSP Methodology and Process
  32. Reuse-Oriented Model Year Architectures for Rapid Prototyping
  33. Support for Model-Year Upgrades in VHDL Test Benches
  34. TRW RASSP Model Year 1 Spread Spectrum Processor
  35. VHDL to Hardware: A TIREP Success Story
  36. VHDL Modeling, Test and Distribution
  37. VHDL-Based Performance Modeling and Virtual Prototyping
  38. Workflow Modeling for Implementing complex, CAD-Based, Design Methodologies

Advanced Technology Laboratories' Path to 4X Improvements

Jeff Pridmore

Abstract

To provide 4X improvements in time-to-market, life-cycle cost, and design quality, Lockheed Martin's Advanced Technology Laboratories is uniquely combining the three elements of the RASSP technology triad -- Methodology, Model Year Architecture, and Infrastructure/Design Environment -- into an integrated, rapid prototyping environment.



Approximate Processing and Incremental Refinement Concepts

J. Winograd , J. Ludwig, H. Nawab , A. Chandrakasan and A. Oppenheim

Abstract

Approximate processing and incremental refinement concepts are needed for applications where it is desirable to provide a systematic tradeoff between the quality of signal processing results and the availability of resources, such as time, bandwidth, memory, and power. We examine the impact of these concepts for three distinct application areas: (1) low-power frequency-selective FIR filtering, (2) real-time time frequency analysis of signals and (3) DCT-based image encoding/decoding. Results from approximate processing of signal data illustrate the practical utility of these types of systems.



Architectures for a RASSP Signal Processor

Fred Shirley, Bob Bassett andJ. P. Letellier

Abstract

This paper summarizes the various elements of a RASSP architecture and describes the architecture selection process that is used to identify a particular architecture for a given system problem. RASSP architectural approaches address general attributes of the signal processor under development, including the communications methods and the type of processing elements used. Architectures of interest span multiple system application areas and lend themselves to periodic model-year upgrades in concert with the RASSP development methodology. Experiences from our Demonstration and Benchmark teams are used to illustrate the alternative results that can arise from different problem constraints. In the case of the Demonstration, the architectural solution consisted of custom interface cards on an industry standard bus using commercially available processors. In the case of the Benchmark the solution consisted of a custom interface and a custom processing card connected through an industry standard bus. Differences in space and power constraints in the two problems account for the two different solutions.



Autocoding in the Lockheed Martin ATL/Camden RASSP Hardware/Software Codesign

Christopher B. Robbins

Abstract

Autocoding provides the Lockheed Martin Advanced Technology Laboratories Camden (LM-ATL Camden) RASSP hardware/software wdesign process the means to rapidly realize implementations of the wdesign software architectures. Management Communications and Control, Inc. (MCCI) autocoding tools automate translation of software architecture specifications to designs and their implementations for functional and detailed hardware/software wsimulation, unit testable modules, and wmplete application load image specifications. The tools support an open application programmers interface to the codesign process. Autocoding tools support the wdesign processes objective of providing a seamless translation of applications from math tool level functional algorithm specifications to target architecture load images. Autocoding technology is directed at reducing the labor content of software design and coding, enabling rapid development of the software elements of application specific signal processing systems.



Automated Generation of Accurate VLSI Behavioral Processor Models for Simulation and Synthesis

Yong-kyu Jung, Vijay K. Madisetti and John W. Hines

Abstract

A new process for automating the creation of Full- Behavioral (FBM) and Instruction Set Architecture (ISA) models in VHDL for complex processors and components is described, with results from the automation of a PowerPC 601 described in some detail. A number of advantages to this approach are described together with its impact on the hardware/software codesign and system prototyping processes.



Collaborative VHDL Modeling within the RASSP Program Demonstration Project

Ray Dreiling and Paul Kalutkiewicz

Abstract

This paper describes experiences and lessons learned in executing a top down VHDL based design with a distributed multi-company team. The Sanders RASSP Demonstration team used a virtual company concept for developing a VHDL description of an IRST (Infrared Search and Track) processor This work spanned three companies across the county and involved all aspects of the project, i.e. design, analysis, fabrication, and test. This paper describes the communications and coordination methods used across the virtual company including the processes to support a distributed design database, source code coherence across multiple networks, and secure communications. The result of these efforts were that the team was able to complete all its modeling and designs without a single collocated design review and with successful completion of all the hardware development.



Creating a "Plug and Play" Architecture to Support Hierarchical Virtual Enterprise Design Process Management

Scott Majdecki and John Purchase

Abstract

Supporting a "Plug & Play" architecture that allows the users choice of tools in the Workflow Manager area is as important as in the individual tool areas. However, like integrating EDA tools, designing an environment that would allow multiple workflow management applications interoperate poses some challenge. This paper will discuss an architecture that will allow a hierarchical design process to be successfully managed by multiple workflow manager applications in a Virtual Enterprise Environment. It will use the Module Design workflow as an example to illustrate an implementation of this architecture.



Developing Re-useable Performance Models for Rapid Evaluation of Computer Architectures Running DSP Algorithms

Hormazd P. Commkkat, Dr. F. Gail Gray, Dr. James R Armstrong and Dr. Geoffrey A Frank

Abstract

This paper discusses the issues in the construction and use of multiprocessor performance models based on an externally developed library of component descriptions. The component models used were initially developed by another RASP contractor (Honeywell), but they were upgraded modified and customized to suit our needs. We compare the simulation times for the VHDL models of the three different architectures, all running a variation of the synthetic aperture radar benchmark provided by MT Lincoln Labs. The three architectures studied are: 1) a common global bus, 2) a two level bus hierarchy with multiple local busses and a single global bus, and 3) a system of local busses connected via a crossbar interconnection. Variations in key performance characteristics of the systems that are used for tradeoff decisions including component utilization, algorithm latency, and algorithm throughput for the three architectures running the same algorithm are recorded.



Environment and Tools for an Integrated RDE (ENTIRE)

Rick Ong, Rob Costantino and Rodger Philips

Abstract

This paper describes the features of Release 1.0 of the RASSP Design Environment (RDE). This release is the fourth internal build of the RDE. The RDE development effort utilized previously developed RDE utilities in conjunction with the RASSP process of software development. This release of the RDE contains a common set of core services for use in a wide range of applications: common desktop, automatic metrics collection, metrics analysis, reuse utility, technical review utility, problem reporting utility, log utility, and remote data access. The RDE is being tailored for use on our Model Year I demonstration to include domain specific tools, translators, libraries, and process support. A preliminary list of applications for the Demonstration are given in the paper. The fully populated and tailored design environment is described as the ENTlRE concept (ENvironment and Tools for an Integrated RDE).



Evaluating Distributed Multiprocessor Designs

Todd Steeves, Fred Rose, Todd Carpenter, John Shackleton and Otto von der Hoff

Abstract

Successful multiprocessor system design for complex real-time embedded applications requires powerful and comprehensive, yet cost-effective, productive, and maintainable modeling. The multi-disciplinary, VHDL-based modeling library developed by the Honeywell Technology Center, and enhanced for the RASSP community, places heavy emphasis on multiprocessing and distributed communications. These models focus on detailed hardware performance analysis along with multiple abstraction levels for software representation and evaluation.



A Framework for the Development of Hybrid Models

Moshe Meyassed, Robert McGraw, James Aylor, Robert Klenke, Ronald Williams, Fred Rose and John Shackleton

Abstract

Rapid prototyping of complex digital systems requires a well defined design flow. A typical top-down design flow starts with a construction of a performance model of the system under design, which helps in making architectural design decisions. Unless this model is used for later phases of the design process, a model continuity problem exists. This problem results from having to model and simulate systems using different design environments for different levels of design detail. Most of the levels of the design process do not exhibit the model continuity problem. However, this problem is prevalent between the performance and functional modeling levels. The work presented here allows for the true step-wise refinement of a performance (uninterpreted) model into a functional or behavioral (interpreted) model. The critical hurdle to the realization of this methodology is the ability to do hybrid modeling. Hybrid modeling is the capability of mixing high-level performance constructs and functional components in a common analysis environment. Hybrid modeling supports the model refinement design flow by providing an interface to bridge the information gap between performance models and behavioral implementations.


Hardware/Software Codesign in the Lockheed Martin Advanced Technology Laboratories RASSP Program

W. Bernard Schaming and Arnold D. Bard

Abstract

A major goal of the RASSP program is to change the way signal processors are designed. One of the key elements in changing the design process is establishing a methodology and an implementation of that methodology that embraces the notion of hardware/software (HW/SW) codesign. The objective of this paper is to define HW/SW codesign and describe the overall codesign process as they are implemented in Lockheed Martin Advanced Technology Laboratories' RASSP program.



Implementation of the RASSP SAR Benchmark on the Intel Paragon

Curtis P. Brown,Richard A. Games and John J. Vaccaro

Abstract

A software design process for mapping real-time applications onto massively parallel processors is described. The design methodology incorporates a software test bench used to evaluate the level of real-time performance the processing nodes are capable of delivering. The final integration step maintains the simple test bench interfaces and reduces the complexity of integrating the components to satisfy the timing requirements of the overall application. The process is applied to implement the RASSP SAR benchmark on an Intel Paragon. The initial implementation uses 12 Paragon GP nodes for a single polarization. Under OSF/1 this 12 node configuration satisfies all the real-time requirements. Under SUNMOS, a streamlined high performance operating system available on the Paragon, the throughput improves significantly with sustained processor utilization approaching 40%. A projected implementation of the RASSP SAR benchmark on the Embedded Touchstone suggests that all three polarizations can be processed (including I/O) using 14 out of its 16 MP nodes.



Integration of DFT into RASSP

John Evans, Richard M. Sedmak and Patrick McHugh

Abstract

This paper describes the integration of Design-For-Testability (DFT) processes, tools, and methodologies into the RASSP System, which consists of the Design Environment and Enterprise System. The blueprint for the DFT developments is the DFT Methodology, which is highly automated, hierarchical, and spans the entire life cycle, contributing significantly to the RASSP goals of 4x improvement in cycle time, design quality, and life cycle costs. Key concepts of the DFT methodology are covered. A preferred testability architecture that encompasses embedded test resources (BIST), external test resources (ATE), and testbenches for design verification is described. Integration of the DFT methodology and testability architecture into the RASSP system is covered. The paper concludes with a discussion of the contribution of DFT to meeting the RASSP goals.



Integrated Process Control and Data Management in RASSP Enterprise Systems

John Welsh, Biju Kalathil, Bipin Chadha, Mary Catherine Tuck, William Selvidge, Elisa Finnie and Arne Bard

Abstract

The RASSP Enterprise System provides key automation support for teams of signal processing/electrical design engineers in the execution of complex development projects. As a result, the system facilitates greatly improved productivity, as well as efficient program control and orderly management of design configurations. Core concepts of the RASSP Enterprise System include integration of tools and tool frameworks into an enterprise environment; program execution control through workflows; integrated data management functions; concurrent engineering team support; and integration of design engineering and manufacturing. This paper presents a strategy for the use of the RASSP environment, methodology/workflows, and information models to improve efficiency in task execution and information management on signal processor development projects.



Integrated Reliability and Performance Evaluation using Information Flow Models

Ramesh Rao, Barry W. Johnson, Ronald D. Williams and James H. Aylor

Abstract

Information flow models provide a natural and convenient representation of the structure, function, and data/control flow of digital electronic systems at a high level of abstraction and interpretation. Design engineers prefer to use such models for performance trade-offs and analysis. Adding the capability to model and analyze the dependability characteristics of a system within such a paradigm allows design engineers, as opposed to reliability engineers, to perform dependability analysis and trade-offs at the early stages of the design process. A design environment supporting such a philosophy would bring dependability analysis and trade-offs into the mainstream of the design process. The Advanced Design Environment Prototype Tool (ADEPT) is based on such a philosophy. This paper presents an overview of ADEPT with a focus on the dependability modeling paradigm. The various solution paths for the evaluation of dependability metrics are presented. The paper will demonstrate the capabilities of ADEPT by modeling the controller for a jet engine magnetic bearing.



KINDLING: A RASSP Application Case Study

Robert H. Paulson

Abstract

The KINDLING case study is applying RASSP processes to develop algorithmic and performance models of an existing subsystem in order to explore architectural options for in-creased performance and functionality. This subsystem is common to several projects in the Lockheed-Martin Information Processing Systems (IPS) business area. Its current implementation consists of a VME board with either 5 or 24 DSP32C processors and classified soft-ware which runs on these processors. This paper will describe the accomplishments to date and plans for the remainder of the project. Requirements and initial high level design have been captured with RDD-100; a functional model of the signal processing algorithms is currently being created using SPW (Signal Processing Work System) and GEDAE (Graphical Entry Distributed Application Environment); and performance models in VHDL for the entire subsystem will be created. Architectural variations to be considered in the case study include processing a signal with one versus multiple DSPs and varying the DSP processor: SHARC, TMS320C40, Mercury architecture with i86Os or SHARCs, and DSP32C are options. Delay parameters for software blocks in the performance models will be determined us-ing actual processors, development tools, or ISA models for each processor option. Metrics from performance models will be compared against performance metrics from previous implementations. An additional goal is to investigate automated code generation, loading, and benchmarking in target systems using GEDAE. GEDAE already supports Mercury Raceway bus systems with i860processors, and soon will be ported to C40 SPOX systems. Process metrics will be recorded during KINDLING and compared with process metrics from the current implementations. The study will also evaluate the applicability of RASSP tools to this task, especially in a classified environment; make recommendations; and suggest improvements.



Lockheed Martin Advanced Technology Laboratories RASSP Second Year Overview

James E. Saultz

Abstract

The goal of the Prototyping of Application-Specific Signal Processors (RASSP) program is to improve by at least a factor of four the cost and time needed to develop and manufacture signal processors. The approach to reaching the program's goalis based on three technology thrusts; methodology, Model Year Architecture, and infrastructure (Enterprise). Using this triad of technology thrusts, Lockheed Martin Advanced Technology Laboratories' (ATL) RASSP team composed of an alliance of companies has implemented the first baseline RASSP system, which represents a significant advance over today's state-of-the-art. The methodology and tools have been used to demonstrate cost and design cycle improvements on the benchmark virtual prototype (VP) and have resulted in the development of a hardware/software system that demonstrated first-pass success. Additional developments being performed during the last two years of the program will provide further benefits, enabling demonstration of 4X improvements in cost and time to market. This paper provides an overview/update of the progress since the 1994 Annual Review.



Managing the RASSP Virtual Enterprise

Bruce Bullock, Ken Streeter, George Muncaster, Mark Hoffman and Mike McCollough

Abstract

The Lockheed-Sanders RASSP development team is composed of four companies: Sanders, Motorola, Hughes, and ISX. These four geographically disparate companies work together daily across all facets of the RASSP program including: Program Management; System Engineering; RASSP Design Environment (RDE) Development; Process Development; Demonstration Work; Benchmark Efforts; and product and process Proliferation. Each major program element involves collaboration by all team members. The associated work products are the result of highly integrated activities. How is this successfully accomplished? continuing RDE development. This paper details the methodologies, tools, and standards that enable the Virtual Corporation to operate successfully and illustrates their use and benefits to its members. An area not addressed in this paper but recognized by the RASSP team as a key issue in the instantiation of a virtual enterprise are the limits, controls, and restrictions involved in: This paper describes a set of tools, methodologies, and standards developed and utilized by the team to enable the establishment and operation of the RASSP virtual corporation.



Projecting RASSP Benefits

James C. Anderson

Abstract

Benefits of a RASSP-based design approach have been identified as a result of benchmarks defined and evaluated by MIT Lincoln Laboratory. One benchmark task, development of a synthetic aperture radar signal processor, serves to illustrate the effects of design trade-offs and tool usage on acquisition cost, development time and life cycle cost. Parametric cost estimation tools are used to predict the benefits of a RASSP-based design approach relative to standard practice, and one example demonstrates a factor of four cost reduction.


Rapid Design and Exploration of Signal Processing System Using a VHDL Model Generator Based Paradigm

Scott R. Powell and Thomas M. Cesear

Abstract

An upgradable design methodology is described for the rapid design and implementation of ASIC-based embedded DSP systems; with results from a detailed design example. The design methodology is based around a library of VHDL model generators which create application-specific simulatable and synthesizable VHDL models of DSP algorithms. The VHDL model generators are extensively parameterized to encapsulate microarchitectural design expertise and elevate it to higher level design activities where there is a greater amount of leverage. Generators provide a cost-effective means to explore the area, speed, and power tradeoffs of algorithmic, architectural, and microarchitectural design alternatives at very early stages in the design process. An expanding library of application-specific, parameterized VHDL model generators is described which is oriented toward common signal processing functions. The library is hierarchical with model complexity ranging from single arithmetic operators to large core functions to entire VLSI chips. The VHDL generator library has been used to design several DSP ASICs; a specific design example is presented demonstrating a 3 to 5 times reduction in overall design time from system specification to mask-level layout.



Rapid Prototyping of Application-Specific Signal Processors: Educator/Facilitator Current Practice (1993) Model and Challenges

Vijay K. Madisetti, Jack Corley and Gary A. Shaw

Abstract

The Rapid Prototyping of Application-Specific Signal Processors (RASSP) project of the US Department of Defense (ARPA and Tri-Services) targets a 4X improvement in the design, prototyping, manufacturing, and support processes (relative to current practice). The authors present a current practice (circa 1993)" model for the design and prototyping of application-specific signal processors developed as part of the RASSP Educator/Facilitator (E&F) Program. A number of limitations in current design practice are highlighted together with challenges faced by candidate solutions. The E&F Program proposes that this model be used as a baseline in the evaluation of newer RASSP prototyping methodologies and processes.


Rapid Prototyping Applied to Underwater Acoustic Modem Research & Development

Paul D. Fiore, Eric Will, Geoffrey Edelson and David Herold

Abstract

Rapid prototyping is well suited to address the signal processing design tasks encountered in research applications. The applications differ from defense or industrial applications in that the signal processing requirements are either evolving or poorly understood. In the field of underwater acoustic (I/WA) communications, practical solutions, and thus their implementations, are still maturing. Sanders and WHOI are using the development of an improved UWA modem to exercise the processes and environments being developed at Sanders under the RASSP program. This paper represents an update to our work previously published. In this paper we provide an overview of the signal processing requirements for UWA communications. We outline the impact of these requirements on system design and describe the application of RASSP processes to system architecture specification. Finally, several potential paths for future development are described, both for system upgrades and for new applications. key aspects of the RASSP process will be evaluated upon completion of this project.



RASSP Benchmark 1: Virtual Prototyping of a Synthetic Aperture Radar Processor

Eric A. Rundquist, Jr.

Abstract

A virtual prototype of a synthetic aperture radar processor is being created by Lockheed Sanders. It includes VHDL models and Ada application code. The virtual prototype is a behavioral and structural model which matches the hardware architecture which was selected as part of this effort. As part of the RASSP program, we are answering the question: "How can one use VHDL /Ada to validate hardware / software tradeoffs and reduce the life cycle cost of digital systems? " The current status is that the virtual prototype has been completed and we are developing a hardware prototype which will validate the virtual prototype.



RASSP Benchmark-1 and -2: A Preliminary Assessment*

A. H. Anderson, G. S. Downs, G. A. Shaw

Abstract

These two benchmarks required the development of a virtual prototype and a hardware prototype, respectively, of a Synthetic Aperture Radar processor. The two RASSP Developers chose different approaches: one used COTS components on custom boards with a methodology emphasis on detailed VHDL prototyping and board design and one used COTS computer boards with a methodology emphasis on efficient VHDL modeling and automatic code generation. Both efforts are briefly described. A preliminary assessment of Benchmark-1, which has been completed, is offered with emphasis on the experience with VHDL modeling. Based on this assessment, some recommendations for improvement are made.



RASSP Process Management

J. Malley, R. Bassett, M. Falco, G. Bruggemann, G. Muncaster

Abstract

IEEE P-1220 as a basis, and is documented in IDEFO format for dissemination to applications both internal and at Beta sites. As the Sanders’ RASSP Team has developed, evaluated and experimented with the process, descriptors use text, hypertext, and graphics. The success of the process evolution has been demonstrated in its successful application. This paper discusses the Sanders’ Team’s process evolution, optimization. and application during the first two years of the program.



RASSP Program Overview

W. Hood, M. Hoffman, J. Malley, C. Myers, R. Ong, E. Rundquist, L. Scanlan,F. Shirley, D. Uyemura

Abstract

This Sanders-led team of Motorola, Hughes and ISX has met all of the primary RASSP program objectives during the first two years of the program. This paper reviews the goals of the program, and the unique ways in which our team is meeting them. The flexible methodology and design environment are described along with the progress made in creating a standard enterprise framework. The progress of the demonstration and benchmarking effort is detailed, as is the work towards proliferation of the RASSP process. The emphasis on VHDL and Ada-based virtual prototyping and its impact on Model Year Upgrades is discussed. The creation of the Virtual RASSP Corporation and the special Internet communication protocols developed to support the program are reviewed. Accomplishments in each of the program areas are reviewed along with specific goals for the next year of the program. Particular emphasis is placed on our Model Year 0 demonstration in which we designed, fabricated, and tested Infrared Search and Track (IRST) flight hard-ware in less than a year. Comparison of the time and resources required to perform Model Year 0 with a comparable non-RASSP development demonstrates that we have already achieved a factor of more than 2.2 X improvement in development time and development cost.



RASSP: The Road to 4X

Larry Scanlan, Ph.D. and LeRoy Fisher

Abstract

The RASSP Program has ambitious goals: 4X decrease in product development cycle-time, 4X decrease in life-cycle costs and 4X increase in product quality. Reaching these goals requires a map so we can choose the route that leads to the desired destination while avoiding financial mountains too high to climb or technology gaps too wide to jump. With the map we can expend all of our energy navigating the routes that will accelerate our progress. In the sections that follow, the factors that contribute to cycle-time and quality will be identified along with the barriers to change. Next we describe some of the approaches being taken by the Lockheed Sanders RASSP team to take advantage of the enabling factors. We will then summarize the results of a product development task analysis to show how each RASSP process reduces cycle-time and improves quality. Finally, we will assess the progress of the IRST Signal Processor Demonstration Team towards achieving 4X.



RASSP Technology Insertion into the Synthetic Aperture Radar Image Processor Application

Junius Pridgen, Richard Jaffe, William Kline

Abstract

This paper describes the development on RASSP Benchmark 1 and 2 of the synthetic aperture radar (SAR) image processor using the RASSP design environment. The overall process flow developed by Lockheed Martin's Advanced Technology Laboratories, as applied on the SAR processor, is illustrated. Results from using executable specifications; parametric cost estimating tools and VHDL-based performance modeling for architecture tradeoffs; hardware/software codesign; virtual prototyping for architecture verification; software generated by autocode; and VHDL-based, top-down hardware development are shown. This paper discusses the implementation strategy and lessons learned on the RASSP benchmark activities.



RASSP VHDL Modeling Terminology and Taxonomy

Carl Hein, Todd Carpenter, Paul Kalutkiewicz, Vijay Madisetti

Abstract

VHDL modeling taxonomy and terminology conventions are emerging from the on-going efforts of the Terminology Working Group (TWG). Based upon examination and comparison of previously published modeling taxonomies, the working group is evolving a multi-axis taxonomy designed to describe the information content of RASSP model types and abstraction levels and to facilitate selection and construction of interoperable models. The TWG used the taxonomy to concisely refine modeling terms applying to system, hardware, and software models; abstraction levels; structural hierarchies; and modeling paradigms. The refined definitions for several of the modeling terms especially important in RASSP are listed and discussed.

Version 1.0 was presented at the conference. This document has been revised and the most current version is provided. This document may be accessed at anytime from the "Side or Bottom" VHDL Taxonomy button.


Real Time IRST Development Using RASSP Methodology and Process

Michael Vahey , Decker Bushman, Steve Dabell, Ted Ennis, Paul Kalutkiewicz, Win0 Lee, Mike McCollough, Brian Nelson, Kathryn Russell, David Uyemura, and Howard Wanke

Abstract

The development of a real-time infrared search and track (IRST) processing system served as a real-world application of the on-going development of methodologies and processes intended to achieve a four time improvement of cycle time at the end of the four year program. This development is an integral part of the Rapid Prototyping of Application Specific Signal Processors (RASSP) program. Real world applications are used to validate and provide metrics on the effectivity of the methodology and processes. A key part of the development methodology is virtual prototyping, a VHDL technique for validating a hardware and software design before hardware is developed. This paper describes the RASSP virtual prototype and overall results achieved by the demonstration team.



Reuse-Oriented Model Year Architectures for Rapid Prototyping

G. Caracciolo and J. Pridmore

Abstract

The Rapid Prototyping of Application-Specific Signal Processors (RASSP) program is striving to change the way embedded signal processor design is performed, providing >4X improvements in time-to-market, cost, and design quality. These improvements will be achieved using a methodology that stresses hardware and software reuse in conjunction with Model Year Architectures that facilitate reusability and upgradability through open interface standards. This paper will describe a Model Year Architecture approach for the development of cost-effective signal processors can be applied to a wide range of military and commercial applications.



Support for Model-Year Upgrades in VHDL Test Benches

Geoffrey A. Frank, James R. Armstrong, Gail Gray

Abstract

The RASSP model-year concept requires a design system that supports evolving system requirements such as changes in sensors, system platforms, targets, and clutter characteristics. Virtual prototyping with VHDL is seen as an essential part of rapid system design in an environment of changing system requirements. Test bench development is often 50% or more of the effort of building a virtual prototype, so effective and efficient means of upgrading test benches is essential to achieving the RASSP goal of a factor of 4 reduction in design costs and time to market. This paper describes two techniques that the Virginia Tech/Research Triangle Institute team is using to support rapid and correct modification of high level VHDL test benches for virtual prototypes. First, the team has been creating VHDL test benches by converting test benches in other tools into VHDL. Second, the team has been developing automated links between requirements databases, test bench component libraries, and VHDL test benches.



TRW RASSP Model Year 1 Spread Spectrum Processor

Carolyn Kuttner

Abstract

TRW is using the Martin RASSP system to develop a Spread Spectrum Preprocessor (SSPP) targeted for the requirements of the Integrated Sensor System (ISS) and Joint Advanced Strike Technology (JAST) programs. Spread Spectrum processing is used to improve communication in noise and jamming and is a key part of many communication links, including JTIDS (Joint Tactical Data Information Distribution System). EPLRS (Enhanced Position Location Reporting System) , and WNA (Wideband Network Access). Spread Spectrum processing is done early in the signal processing flow and so requires a very high computational bandwidth -- 2-channel JTIDS needs 5 BOPS and four-channel EPLRS requires 50 BOPS. The overall programmatic goals of ISS and JAST are in line with RASSP’s new capabilities -- including reducing life cycle costs with simulation and affordable upgrades to open systems. The specific SSPP requirements are also in line with RASSP, emphasizing physical and logical scalability and flexibility. This paper describes the RASSP process being used status, and lessons learned to date.



VHDL to Hardware: A TIREP Success Story

Ed Woods, Darin York, Gary Hout, John Miles, L.J.Ceder, Charles Rogers, David Broadhead, Louie Kitcoff, Lindsay Skidmore

Abstract

Electronics obsolescence in military systems is one of the most expensive and elusive challenges facing the Department of Defense today. This one issue is costing millions of dollars a year in re-engineering, special orders, volume buys and redesign costs. Development of new techniques, tools and processes for maintenance of legacy systems and integration of new technology in an era of life cycle extension and funding reduction is essential. The Technology Independent Representation of Electronic Products (TIREP) project was funded by the Standard Hardware Acquisition and Reliability Program (SHARP) and the Flexible Computer Integrated Manufacturing (FCIM) program offices to address these repair, reliability, and obsolescence issues in the maintenance of legacy systems. TIREP is a joint effort between the Naval Research Lab (NRL), Washington DC, the Naval Air Warfare Center - Aircraft Division Indianapolis (NAWC-ADI), and the Naval Surface Warfare Center (NSWC) - Crane Division. TIREP has developed a seamless process to cost effectively recreate a form, fit, and function replacement of electronic circuit card assemblies from a digital functional behavioral description such as the Very High Speed Integrated Circuit(VHSIC) Hardware Description Language (VHDL) along with other related standards.



VHDL Modeling, Test, and Distribution

Vincent L. Sanders, Robert B. Reese, Aubrey K. Knight, J. Scott Calhoun

Abstract

This paper presents the ongoing effort of the Mississippi State University / Microsystems Prototyping Laboratory (MSU/MPL) executed as part of the RASSP Technology Base Program. The research underway includes: VHDL model development, VHDL test bench synthesis, and utilization of the World Wide Web (WWW) to document and distribute models, tools, and information.

To date, modeling guide lines have been implemented for PROMs, SRAMs, Dual–Port SRAMs, FIFOs, and PLDs. Successful models have been developed using these guidelines. Guidelines for CPLDs and FPGAs are currently being developed.

A VHDL Test Bench Synthesis Tool has been developed and is scheduled for beta testing in July 1995. This tool automatically generates VHDL test benches and allows for vendor independent testing methodologies of VHDL designs.

The World Wide Web (WWW) is being used to document and release the VHDL models developed. Complete datasheets for sample models have been developed and posted to the Web as an example of hierarchical electronic datasheets.



VHDL-Based Performance Modeling and Virtual Prototyping

Carl Hein and David Nasoff

Abstract

The need for top-down, hierarchical, simulation processes in RASSP has pushed the use of VHDL at levels of abstraction above where it has typically been used. Selection of an appropriate modeling abstraction level permits rapid exploration of many alternative software and hardware solutions while achieving accurate yet efficient simulation. Performance and virtual prototype models facilitate the integrated design and development of the software and the hardware subsystems in a hardware/ software codesign process. In this paper, the Lockheed Martin Advanced Technology Laboratories (ATL) RASSP team will present methods for producing an efficient VHDL-based performance model of a Digital Signal Processor (DSP) system, which is then extended to form a comprehensive virtual prototype of a full multi-processor DSP system that is both timing- and data-faithful. The system models provide early design verification via simulation of software, as partitioned, mapped, and executed on the hardware architecture. This paper will focus on VHDL-specific issues and techniques for this type of very abstract modeling.



Workflow Modeling for Implementing Complex, CAD-Based, Design Methodologies

J. Stavash, J. Wedgwood, M. Forte, W. Selvidge, M.C. Tuck, A. Bard and E. Finnie

Abstract

A specific goal of the Rapid Prototyping of Application-Specific Signal Processors (RASSP) program is to achieve a 4X improvement in design cycle time, cost of the design, and the quality of the final product. In order to achieve these 4X improvements, Lockheed Martin Advanced Technology Laboratories (ATL) is developing design methodologies that make use of concurrent engineering, design object reuse, and the spiral model of development for rapid prototyping. Lockheed Martin ATL, Rockwell International, Intergraph, and Aspect Development are collaborating on developing workflow models to implement these design methodologies, as well as the data management and CAD tool environment to support them.