The Definicon 68020 Coprocessor

Total Page:16

File Type:pdf, Size:1020Kb

The Definicon 68020 Coprocessor PA RT THE HARDWA RE AND OPERATING SYSTEM 1: THE DEFINICON 68020 COPROCESSOR tem" on page 122 for complete dress decoding circuitry is shown in A plug--in board that information.) figure 3 and includes the FPU20 and The DSI-020 is a coprocessor board DEC20 PALs. the HOLD20 bus re­ provides 2--bit (see photo I) that uses the 12.5-mega­ quest arbiter PA L. the 32-bit bus DMA 3 hertz versions of the 68020 CPU and sequencer (74FI6I and DPORT20 computing power and the 6888I FPU (floating-point unit). PA L). the DSACK20 PAL. and the Provision has been made for the ad­ BE20 byte enable PA L (figure 4). The dition of the 688 51 PMMU (paged­ DSACK20 PA Ls main function is to math processing for memory-management unit) when it provide signals to the 68020 to in­ becomes available. The board has I dicate that a memory access cycle is 16--bit machines megabyte of high-speed parity­ complete and whether the size of the checked dynamic RAM that requires transfer was 8. I 6. or 3 2 bits. only one CPU wait state to operate A delay line and associated headers ave you ever wished that correctly. There are two high-speed delay the CPU's address strobe (AS) your personal computer (38.400 bps) serial ports (figure I) and signal long enough for all the de­ had the speed of a VA X an interrupt-driven timer to ease the coding to have been completed be­ or that the number 64K task of porting UNIX. fore the RASIN signal is generated to had never existed? We ll. the Motorola All the facilities of a host IBM PC or initiate the DP8409 dynamic RAM 68020 32-bit microprocessor has the compatible are available to the co­ controller's access cycle. The DP8409 computing power to come close to processor. including networking and (figure 5) directly drives the multi­ fulfillingH both those dreams. BYTE has access to any other peripherals con­ plexed address and RAS lines of the arranged for Definicon Systems to nected to the PC's bus. System calls RAMs. The CAS signal is multiplexed design a special version of its 68020- for full bit-mapped graphics are pro­ with the Byte Enable signals before based scientific microcomputer sys­ vided in the kernel. All BDOS and being applied to the RAM array, tem that allows 68020 hardware and BIOS software interrupts in the PC are shown in figure 6. The refresh counter software development at a price all available to the 68020 programmer. is driven from the CPU's clock oscil­ can afford. (See the text box "How to lator. divided by 128. The DP8409 at­ Get Yo ur DSI-020 Coprocessor Sys- OVERVIEW tempts to do a hidden refresh, but if The DSI-020 board is composed of a this is not possible, a refresh request Trevor Marshall. Christopher jones. and Sigi number of relatively independent (RFIO) will be sent to the HOLD20 Kluger are engineers with Definicon Systems functional blocks. The CPU cluster PA L every 10.24 microseconds. When Inc. They can be contacted at 3 I 324 Via Col­ (figure 2) consists of the 68020. this is acknowledged (with RFSHACK) inas #I08/9, Westlake Village, CA 9I 362. 6888 1. and 688 5 I. The 3 2-bit bus ad- the DP8409 will perform a RAS ONLY refresh cycle. Hidden refreshes occur whenever any device other than the RAM itself is addressed. In particular. accesses to the 68881 will cause a hidden refresh to occur. This considerably reduces any overhead that might otherwise be incurred using this DMA refresh technique. 120 BYTE • JULY J986 BY TREVOR MARSHALL, CHRISTOPHER }ONES, AND SIGI KLUGER Photo I: Tlie Definicon DSI-020 board with 68020 CPU. 68881 FPU. and I megabyte of memory. A 688 51 PMMU is shown with this board althougli it is not commercially available at this time. The parity error signal generated as implement virtual memory. The map­ and performs a memory cycle. The shown in figure 7 is latched and can ping from logical to physical ad­ special coding on the status lines in­ be read by the host PC at bit 0 of the dresses is stored in tables (in memory) dicates. ho·.vever. that the cycle is ac­ BFLAGS port (see table I and figure that the PMMU can search. It has an tually for the coprocessor interface. 4). Pulse the Refresh Inhibit line internal cache so that the most Chip select for the 68881 is derived 1 (RFSHINH) briefly to reset this latch. recently used translation table entries in the FPU20 PAL. are available without the necessity for The major feature of the 68881 is THE 68851 PMMU an additional bus cycle. The PMMU that many frequently used mathemat­ The principal function of the PMMU implements a hierarchical protection ical functions have been microcoded is the translation of logical addresses mechanism with up to eight access into the silicon. This considerably from the software running in the CPU levels. It also provides a breakpoint enhances the performance of the to physical addresses within the range acknowledge facility to support the 68020/68881 combination on engi­ of the available hardware facilities. 68020 breakpoint instructions. neering and scientific computations. This allows programs to run at ad­ The 688 51 internally decodes its It also accounts for its deceptively dresses that are not available in own chip select function. high performance on some function­ physical memory. An operating sys­ sensitive benchmarks. such as the tem that swaps infrequently used THE 6888 I COPROCESSOR Savage benchmark (Dr. Dobb's journal. memory blocks to disk and so en­ The 68881 is the first high-speed float­ March 1984. page 92). larges the effective addressing range ing-point accelerator from Motorola. of programs is called a virtual mem­ It is addressed when the 68020 places THE HOST PC BUS INTERFACE ory system (VMS). Although Definicon Ill on its status lines (FCO. FCI. and The IBM PC XT bus was designed as has written a VMS for the DSI-32. the FC2). 0010 on lines Al9 through Al6. a single master bus. so there is no kernel for the DSI-020 does not yet and 001 on lines Al5 through Al3 (continued) PHOTOG RAPHED BY PA UL AV IS JULY 1986 • BYTE 121 DEFINICON 68020 COPROCESSOR dynamic RAM refresh cycle. This (August 1985 BYTE. page 120) we The 8088 is used forces a coprocessor system such as found that the total replacement of the DSI-020 either to replace the 8088 the 8088 was relatively inefficient and to relieve the 68020 or to work in close cooperation with that it is capable of relieving the it. coprocessor of many of the tedious coprocessor of With the design of the PC AT. pro­ disk and console 1/0 tasks. The vision was made for a slave processor DSI-020 follows the same design to control the bus (via the ORO and criteria. The M5-DOS system on either many disk and console MASTER signals). but a limitation of a PC AT or a PC XT is used to pro­ 1 5 microseconds remains on the total vide file 1/0. networking. and other pe­ tasks. length of time a slave can keep con­ ripheral tasks. allowing the 68020 I/0 trol before memory refresh fails! This kernel to concentrate on the things it limit obviously makes it very difficult does best. The kernel provides the in­ method for a slave coprocessor to to provide safe direct disk or other terface and translation between the take control of the bus. tell the 8088 stream 110 tasks even using these UNIX-derived 68020 system software CPU to go to sleep. and drive the newer capabilities. It is for these and the facilities provided by M5-DOS peripherals directly. The XT relies on reasons that the DSI-020 is designed itself. We allowed the optional use of the 8088's DMA controller to take to transfer data 1 byte (or word) at a the 16-bit data transfer facility of the back control of the bus every . 15 time to the host environment. new AT bus. The majority of the im­ microseconds or so to perform a During the design of the OSI- 3 2 provement resulting from the use of How TO GET YO UR 051�020 COPROCESSOR SYSTEM Definicon Systems will provide . the a system with only floppy disks. a fixed Living Software BASIC to C Con- . elements of the DSI-020 system at the disk is essential for meaningful pro­ verter: $299 following special prices for BYTE gram development. Includes run-time package compiled readers. for compatibility with either SVS C Note: These special prices have been DSI-020 basic hardware: $99 5 or LLL C provided by Definicon. reducing its I megabyte of RAM. 12.5-MHz 68020. margins to a bare minimum. In par­ and !2.5-MHz 68881 SVS C: $349 ticular. no margin has been allowed for No parity chips or serial ports Kernighan and Ritchie definition with accounting overhead and so no purchase MS-DOS interface software. simplified most UNIX extensions orders can be accepted for these special assembler. linker. and assembly-level BYTE products. debugger Te rms of payment are VISA/Master­ The DSI-020 hardware is supplied fully LLL C: $349 Card/American Express only. There is assembled and tested. Kernighan and Ritchie definition with a 30-day. no-questions-asked. money­ most UNIX extensions back guarantee. Goods must be re­ Optional hardware upgrade: $99 turned in "as new" condition in original Chips for the two serial ports.
Recommended publications
  • Lecture 1: Course Introduction G Course Organization G Historical Overview G Computer Organization G Why the MC68000? G Why Assembly Language?
    Lecture 1: Course introduction g Course organization g Historical overview g Computer organization g Why the MC68000? g Why assembly language? Microprocessor-based System Design 1 Ricardo Gutierrez-Osuna Wright State University Course organization g Grading Instructor n Exams Ricardo Gutierrez-Osuna g 1 midterm and 1 final Office: 401 Russ n Homework Tel:775-5120 g 4 problem sets (not graded) [email protected] n Quizzes http://www.cs.wright.edu/~rgutier g Biweekly Office hours: TBA n Laboratories g 5 Labs Teaching Assistant g Grading scheme Mohammed Tabrez Office: 339 Russ [email protected] Weight (%) Office hours: TBA Quizes 20 Laboratory 40 Midterm 20 Final Exam 20 Microprocessor-based System Design 2 Ricardo Gutierrez-Osuna Wright State University Course outline g Module I: Programming (8 lectures) g MC68000 architecture (2) g Assembly language (5) n Instruction and addressing modes (2) n Program control (1) n Subroutines (2) g C language (1) g Module II: Peripherals (9) g Exception processing (1) g Devices (6) n PI/T timer (2) n PI/T parallel port (2) n DUART serial port (1) g Memory and I/O interface (1) g Address decoding (2) Microprocessor-based System Design 3 Ricardo Gutierrez-Osuna Wright State University Brief history of computers GENERATION FEATURES MILESTONES YEAR NOTES Asia Minor, Abacus 3000BC Only replaced by paper and pencil Mech., Blaise Pascal, Pascaline 1642 Decimal addition (8 decimal figs) Early machines Electro- Charles Babbage Differential Engine 1823 Steam powered (3000BC-1945) mech. Herman Hollerith,
    [Show full text]
  • A Modular Soft Processor Core in VHDL
    A Modular Soft Processor Core in VHDL Jack Whitham 2002-2003 This is a Third Year project submitted for the degree of MEng in the Department of Computer Science at the University of York. The project will attempt to demonstrate that a modular soft processor core can be designed and implemented on an FPGA, and that the core can be optimised to run a particular embedded application using a minimal amount of FPGA space. The word count of this project (as counted by the Unix wc command after detex was run on the LaTeX source) is 33647 words. This includes all text in the main report and Appendices A, B and C. Excluding source code, the project is 70 pages in length. i Contents I. Introduction 1 1. Background and Literature 1 1.1. Soft Processor Cores . 1 1.2. A Field Programmable Gate Array . 1 1.3. VHSIC Hardware Definition Language (VHDL) . 2 1.4. The Motorola 68020 . 2 II. High-level Project Decisions 3 2. Should the design be based on an existing one? 3 3. Which processor should the soft core be based upon? 3 4. Which processor should be chosen? 3 5. Restating the aims of the project in terms of the chosen processor 4 III. Modular Processor Design Decisions 4 6. Processor Design 4 6.1. Alternatives to a complete processor implementation . 4 6.2. A real processor . 5 6.3. Instruction Decoder and Control Logic . 5 6.4. Arithmetic and Logic Unit (ALU) . 7 6.5. Register File . 7 6.6. Links between Components .
    [Show full text]
  • Introduction of Microprocessor
    Introduction of Microprocessor A Microprocessor is an important part of a computer architecture without which you will not be able to perform anything on your computer. It is a programmable device that takes in input, performs some arithmetic and logical operations over it and produces desired output. In simple words, a Microprocessor is a digital device on a chip which can fetch instruction from memory, decode and execute them and give results. Basics of Microprocessor – A Microprocessor takes a bunch of instructions in machine language and executes them, telling the processor what it has to do. Microprocessor performs three basic things while executing the instruction: 1. It performs some basic operations like addition, subtraction, multiplication, division and some logical operations using its Arithmetic and Logical Unit (ALU). New Microprocessors also perform operations on floating point numbers also. 2. Data in a Microprocessor can move from one location to another. 3. It has a Program Counter (PC) register that stores the address of the next instruction based on the value of PC, Microprocessor jumps from one location to another and takes decision. A typical Microprocessor structure looks like this. Clock Speed of different Microprocessor: 16-bit Microprocessor – 8086: 4.7MHz, 8MHz, 10MHz 8088: more than 5MHz 80186/80188: 6MHz 80286: 8MHz 32-bit Microprocessor – INTEL 80386: 16MHz to 33MHz INTEL 80486: 16MHz to 100MHz PENTIUM: 66MHz 64-bit Microprocessor – INTEL CORE-2: 1.2GHz to 3GHz INTEL i7: 66GHz to 3.33GHz INTEL i5: 2.4GHz to 3.6GHz INTEL i3: 2.93GHz to 3.33GHz We do not have any 128-bit Microprocessor in work at present one among the reasons for this is that we are a long way from exhausting the 64 bit address space itself, we use it a constant rate of roughly 2 bits every 3 years.
    [Show full text]
  • A Manual for the Assemblerߤ Rob Pike Lucent Technologies, Bell Labs
    A Manual for the Assembler Rob Pike Lucent Technologies, Bell Labs Machines There is an assembler for each of the MIPS, SPARC, Intel 386, ARM, PowerPC, Motorola 68010, and Motorola 68020. The 68020 assembler, 2a, is the oldest and in many ways the prototype. The assemblers are really just variations of a single program: they share many properties such as left-to-right assignment order for instruction operands and the synthesis of macro instructions such as MOVE to hide the peculiarities of the load and store structure of the machines. To keep things concrete, the first part of this manual is specifically about the 68020. At the end is a description of the differences among the other assemblers. Registers All pre-defined symbols in the assembler are upper-case. Data registers are R0 through R7; address registers are A0 through A7; floating-point registers are F0 through F7. A pointer in A6 is used by the C compiler to point to data, enabling short addresses to be used more often. The value of A6 is constant and must be set during C program initialization to the address of the externally-defined symbol a6base. The following hardware registers are defined in the assembler; their meaning should be obvious given a 68020 manual: CAAR, CACR, CCR, DFC, ISP, MSP, SFC, SR, USP, and VBR. The assembler also defines several pseudo-registers that manipulate the stack: FP, SP, and TOS. FP is the frame pointer, so 0(FP) is the first argument, 4(FP) is the second, and so on. SP is the local stack pointer, where automatic variables are held (SP is a pseudo-register only on the 68020); 0(SP) is the first automatic, and so on as with FP.
    [Show full text]
  • Linux User Group HOWTO Linux User Group HOWTO Table of Contents Linux User Group HOWTO
    Linux User Group HOWTO Linux User Group HOWTO Table of Contents Linux User Group HOWTO..............................................................................................................................1 Rick Moen...............................................................................................................................................1 1. Introduction..........................................................................................................................................1 2. What is a GNU/Linux user group?......................................................................................................1 3. What LUGs exist?................................................................................................................................1 4. What does a LUG do?..........................................................................................................................1 5. LUG activities......................................................................................................................................1 6. Practical suggestions............................................................................................................................1 7. Legal and political issues.....................................................................................................................2 8. About this document............................................................................................................................2 1. Introduction..........................................................................................................................................2
    [Show full text]
  • Instruction Manual TMS 202 68020 & 68EC020 Microprocessor Support
    Instruction Manual TMS 202 68020 & 68EC020 Microprocessor Support 070-9820-00 There are no current European directives that apply to this product. This product provides cable and test lead connections to a test object of electronic measuring and test equipment. Warning The servicing instructions are for use by qualified personnel only. To avoid personal injury, do not perform any servicing unless you are qualified to do so. Refer to all safety summaries prior to performing service. Copyright E Tektronix, Inc. All rights reserved. Licensed software products are owned by Tektronix or its suppliers and are protected by United States copyright laws and international treaty provisions. Use, duplication, or disclosure by the Government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at DFARS 252.227-7013, or subparagraphs (c)(1) and (2) of the Commercial Computer Software – Restricted Rights clause at FAR 52.227-19, as applicable. Tektronix products are covered by U.S. and foreign patents, issued and pending. Information in this publication supercedes that in all previously published material. Specifications and price change privileges reserved. Printed in the U.S.A. Tektronix, Inc., P.O. Box 1000, Wilsonville, OR 97070–1000 TEKTRONIX and TEK are registered trademarks of Tektronix, Inc. SOFTWARE WARRANTY Tektronix warrants that the media on which this software product is furnished and the encoding of the programs on the media will be free from defects in materials and workmanship for a period of three (3) months from the date of shipment. If a medium or encoding proves defective during the warranty period, Tektronix will provide a replacement in exchange for the defective medium.
    [Show full text]
  • 1. Types of Computers Contents
    1. Types of Computers Contents 1 Classes of computers 1 1.1 Classes by size ............................................. 1 1.1.1 Microcomputers (personal computers) ............................ 1 1.1.2 Minicomputers (midrange computers) ............................ 1 1.1.3 Mainframe computers ..................................... 1 1.1.4 Supercomputers ........................................ 1 1.2 Classes by function .......................................... 2 1.2.1 Servers ............................................ 2 1.2.2 Workstations ......................................... 2 1.2.3 Information appliances .................................... 2 1.2.4 Embedded computers ..................................... 2 1.3 See also ................................................ 2 1.4 References .............................................. 2 1.5 External links ............................................. 2 2 List of computer size categories 3 2.1 Supercomputers ............................................ 3 2.2 Mainframe computers ........................................ 3 2.3 Minicomputers ............................................ 3 2.4 Microcomputers ........................................... 3 2.5 Mobile computers ........................................... 3 2.6 Others ................................................. 4 2.7 Distinctive marks ........................................... 4 2.8 Categories ............................................... 4 2.9 See also ................................................ 4 2.10 References
    [Show full text]
  • Initial Experience with the Intel I860 Microprocessor
    Version 2.1/UPR-0184E Initial Experience with the Intel i860 Microprocessor L.D. Gladney, P.T. Keener, N.S. Lockyer, K.J. Ragan David Rittenhouse Laboratory, University of Pennsylvania, Philadelphia, PA, 19104 J.G. Heinrich, K.T. McDonald Joseph Henry Laboratories, Princeton University, Princeton, NJ, 08544 March 12, 1990 Abstract Scalar performance of the 33Mhz Intel i860 microprocessor is presented. Standard High Energy Physics programs have been ported to the i860 environment. Benchmarks have been performed and the problems encountered are discussed. Testing of the vector features of the i860 has recently started. Introduction The Intel i8601 microprocessor utilizes VLSI hardware design (> 106 transistors) with 64-bit RISC architecture to provide high computational throughput. This is the first in a family of processors being developed by Intel that is expected to achieve 200 MIPS per processor by the year 1998. 2 Our interest in the i860 is enhanced by an ongoing project at Intel Scientific Computers to produce a processor-farm architecture using the i860 with 200 Mbyte/sec data paths between nodes.3 Such an architecture, based on an array of high performance processors interconnected with high bandwidth data paths, is well suited to both online and offline computing needs of the next generation4 of High Energy Physics (HEP) experiments. 1i860, 386, and 387 are trademarks of Intel Corporation. 2L.D. Gladney et al., Proposal to the SSC Laboratory for Research and Development of a Parallel Computing Farm (Oct. 1989). 3see, for example, Federal Computer Week, Vol. 3, No. 15 (April 10, 1989). 4 The BCD Collaboration, An Intermediate and Low pT Dectector for the SSC, SSC-240.
    [Show full text]
  • F Re Esca Le S Em Ic Onduc to R, I Nc .
    Freescale Semiconductor, Inc. WHITE PAPER: Motorola® ColdFire™ VL RISC PROCESSORS Enabling the Next Generation of Advanced Consumer Electronics . c August 1996 n I , r o t c u d n o c i m e S e l a c s e e r F Motorola ColdFire VL RISC Processors 1 For More Information On This Product, Go to: www.freescale.com Freescale Semiconductor, Inc. Table of Contents Overview ....................................................................... 3 The Evolving Embedded Market for Advanced Consumer Electronics ....4 The Home-Based Digital Office .........................................................5 . Mass Storage to Handle the Digital Tidal Wave .......................................5 . c The PC/TV and a Wired Home ..........................................................6 n I Choosing the Right Processor ...........................................................6 , r o The Changing Buyer-Seller Relationship .................................... 7 t c New Product and Service Models .......................................................8 u d ColdFire -- An Innovative Platform for n a New Generation of Advanced Consumer Electronics ..................... 8 o c The Legacy of the Motorola 68000......................................................9 i m Architectural Overview ................................................................. 10 e S The VL RISC Advantage ............................................................... 11 e l Family Overview ........................................................................ 12 a c Conclusion
    [Show full text]
  • Plan 9 C Compilers ߤ
    Plan 9 C Compilers Ken Thompson [email protected] ABSTRACT This paper describes the overall structure and function of the Plan 9 C compilers. A more detailed implementation document for any one of the compilers is yet to be written. 1. Introduction There are many compilers in the series. Eight of the compilers (MIPS 3000, SPARC, Intel 386, AMD64, Power PC, ARM, DEC Alpha, and Motorola 68020) are considered active and are used to compile current versions of Plan 9 or Inferno. Several others (Motorola 68000, Intel 960, AMD 29000) have had only limited use, such as to program peripherals or experimental devices. 2. Structure The compiler is a single program that produces an object file. Combined in the compiler are the traditional roles of preprocessor, lexical analyzer, parser, code generator, local optimizer, and first half of the assem- bler. The object files are binary forms of assembly language, similar to what might be passed between the first and second passes of an assembler. Object files and libraries are combined by a loader program to produce the executable binary. The loader combines the roles of second half of the assembler, global optimizer, and loader. The names of the compli- ers, loaders, and assemblers are as follows: SPARC kc kl ka PowerPC qc ql qa MIPS vc vl va Motorola 68000 1c 1l 1a Motorola 68020 2c 2l 2a ARM 5c 5l 5a AMD64 6c 6l 6a DEC Alpha 7c 7l 7a Intel 386 8c 8l 8a AMD 29000 9c 9l 9a There is a further breakdown in the source of the compilers into object-independent and object-dependent parts.
    [Show full text]
  • HP E2426A/B Analysis Probe for Motorola 68020/EC020 User's Guide
    User’s Guide Publication Number E2426-97001 March 1998 For Safety Information, Warranties, and Regulatory Information, see the pages at the end of this manual. Copyright Hewlett-Packard Company 1993 - 1998 All Rights Reserved. HP E2426A/B Analysis Probe for Motorola 68020/EC020 The HP E2426A/B Analysis Probe — At a Glance The HP E2426A Analysis Probe provides a complete interface for state or timing analysis between of the supported 68020 microprocessors listed below and HP logic analyzers. The HP E2426B Analysis Probe provides a complete interface for the supported 68EC020 microprocessors. The supported logic analyzers are listed in chapter 1. Supported Microprocessors Microprocessor Package Ordering Information 68020 114-pin PGA E2426A 68020 132-pin QFP E2426A option 1CC 68EC020 100-pin PGA E2426B 68EC020 132-pin QFP E2426B option 1CC The analysis probe provides the physical connection between the target microprocessor and the logic analyzer. The configuration software on the enclosed disks set up the logic analyzer for compatibility with the analysis probe. The inverse assemblers on the disks let you obtain displays of the 68020 data in 68020 assembly language mnemonics. If you are using the analysis probe with the HP 16600 or HP 16700 series logic analysis systems, you only need this manual as a reference. The HP 16600 and 16700 series contain a Setup Assistant, which guides you through the connection and configuration process using on-screen dialog windows. For an overview of Setup Assistant, refer to Chapter 1, "Setup Assistant." For more information on the logic analyzers or microprocessor, refer to the appropriate reference manuals for those products.
    [Show full text]
  • Real-Time Radarimage
    Real-Time Radar Image Understanding: A Machine­ Intelligence Approach Ann Marie Aull, Robert A. Gabel, and Thomas J. Goblick II Machine intelligence (MI) techniques have been combined with conventional signal processing and image processing techniques to build a software package that automatically recognizes reentry vehicles from a sequence of radar images. This software package, called ROME (Radar Object Modeling Environment), takes as its input a time sequence ofrange-Doppler images ofa target produced by the Lincoln Laboratory Discrimination System from raw radar-data samples. ROME then processes thi~ image sequence to extract radar features and track them from image to image. From these feature tracks, ROME then constructs a three-dimensional model ofthe target. The object model derived from the image sequence is then compared to a catalog ofmodels ofknown objects to find the best match. Ifno sufficiently close match is found, the observed object is declared unrecognized. The object-model catalog is constructed by adding new models derived from radar data that do not match any model already in the catalog. Thus ROME is "trained" to recognize objects by using real radar data from known objects. Because ofthe strong interest in real-time recognition ofreentry vehicles, the entire ROME system was recoded for parallel execution on the MX-l multi­ processor, which was developed in the Machine Intelligence group at Lincoln Laboratory. This machine was designed for MI applications that involve intensive numeric as well as symbolic computation, and that have real-time processing requirements. The final version ofthe ROME object-recognition system, coded in a combination ofparallel Common LISP and C, runs in real time on the 16-node MX-l multiprocessor.
    [Show full text]