Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
On Mon, 18 Dec 1995, Maya Reuveni wrote: > My question is about floor planning for Xilinx X4013. > I am working on a large design to be mapped into a > Xilinx XC4013 device. The design consists of 4 main > blocks; utilization is 55%; and target speed is 33 Mhz! > > Does anyone have any tips for the grafical floor planner? > > ANY responce will be appreciated. i don't have direct experience with the xilinx floorplanner yet (i'll be targeting an xc4013e in the near future), but what i've heard from the engineers here who have used the xilinx graphical floorplanner, is that it is really only useful for highly structured designs (i.e., those in which the functionality is broken up into smaller, logical chunks, which are assembled into a hierarchy, and have nice names attached to them, etc. -- you know what i'm talking about). so, if your design is not already well structured, maybe you should consider doing that first, before getting into floorplanning, if time permits. otherwise, the floorplanner will not buy you much at all. phil ._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._. Phil Sailer -- EMC Corporation -- 508.435.1000, x4477Article: 2501
I have an ORCA FPGA (2C15) that is configured in the asynchronous peripheral mode. I'm interested in different methods for getting the *.bit information to the peripheral FPGA. One method is to parse the *.bit file into an array, and then use the stored array data as input to the FPGA. If anyone has already succeeded in doing this I would be interested in the program/code. -SimonArticle: 2502
In article <chuck.1169690217A@news.connectnet.com>, Charles P. Ohrbom <chuck@aeroastro.com> wrote: >I need to implement a UA(R)T, i.e. the transmitt portion only, in an Actel >PLD. I will use a pin on a MC68332 as the UAR(T) receiver. Does anyone >have a design they could pass along? > >Chuck Ohrbom >AeroAstro >chuck@aeroastro.com The 1991 version of the Actel handbook had a design for a UART in the application notes. I seem to remeber the transmitter as being the simpler of the two sections (just shift the data out with start/stop bits added at the correct rate). Eric http://www.sentex.net/~eric -- Eric Pearson -- Focus Systems -- Waterloo, Ontario ecp@focus-systems.on.ca (519) 746-4918 "We Engineer Innovative Imaging Solutions"Article: 2503
PRESS RELEASE For Immediate Release December 12, 1995 UPCOMING SHORT COURSE ^^^^^^^^^^^^^^^^^^^^^ VLSI DESIGN AND TEST Georgia Institute of Technology March 18-22, 1996 The Georgia Institute of Technology (Atlanta, GA) is offering a continuing education short course in the area of very large-scale integrated (VLSI) circuit design. The lectures focus on the custom and semicustom design, verification, testing, and packaging of digital and mixed-signal integrated circuits. A design-intensive laboratory experience provides participants with hands-on experience designing and verifying a complete mixed-signal VLSI chip. For more information: World-Wide Web: http://www.ee.gatech.edu/academic/conted/VLSI/ Registration or brochure: conted@gatech.edu OR 404-894-2547 AND ALSO http://www.conted.gatech.edu Course content/requirements: steve.deweerth@ece.gatech.eduArticle: 2504
I have several questions related to the use of Altera components. 1) clock enable In a previous thread, the use of clock "gating" in Xilinx XC40xx LCA was under discussion, pointing out towards the use of the dedicated Clock Enable terminal of a CLB/IOB, which is of course very useful and easy to use. My question is now related to the same for Altera-EPLD EPM7000 family. In the data sheet it is clearly shown that the basic register structure has a clock enable signal. Which input of which symbol of which library must be considered to use this clock enabling? I have a synchronous work around, but that needs more resources. 2) Has anyone used the XNF interface writer of the MAX2Plus tool chain. What about AHDL. Common libraries and target independent inputs? TIA. Regards, Jan. -- __________________________ Jan Maris Zaventem, Belgium. jan.maris@ping.beArticle: 2505
I am fishing for trends here... In an effort to kick start my engineering career into existance, I intend to learn one of these HDLs soon. The question is, which one? Industry seems not to hire generalists so it seems best to put the energy into one rather that divide among both. Which HDL would provide the greatest benefit for someone trying to break into logic design? ---- Eric Edwards is, and will always be: eric@exile.org Remember the home hobbyist computer: Born 1975, died April 29, 1994Article: 2506
Hi, Years ago, I've ever got one, about 100 pages document, from their local distributor: OPEN-ABEL Technical Specification P/N:971-1134-001 I think sign up on the "OPEN-ABEL Registration Card" is required. How about contacting to your local distributor, or DATA-I/O product marketing division directly? I hope this helps. In article <JPA.95Dec14111640@hobbes.inesc.pt> jpa@hobbes.inesc.pt (Jose'Pedro Abreu) writes: > I am looking for some pointers regarding OpenABEL. I have the ABEL > manuals from DATA-IO but there little referrence to OpenABEL. > > I've checked the www site from DATA-IO but I couldn't find anything. --- H.Miyauchi --Article: 2507
chuck@aeroastro.com (Charles P. Ohrbom) writes: >Hello, >I need to implement a UA(R)T, i.e. the transmitt portion only, in an Actel >PLD. The german PC-Magacien "C'T" had an article in 1988 #3 Page 155 for both direction. The simplest transmit-part uses 2 HCT 165-Chips [shift-registers] and a clock-generator: D7 D6 D5 D4 D3 D2 D1 D0 Strobe +5V -*--*--*--*--*--*--* | | | | | | | | | | | | | | | | | | | | | | | | M +5V | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A B C D E F G H A B C D E F G H | | | | v *--Ser IN Out --- Ser in Out -------------> ^ | Clock Inh Load Clock Inh Load | | | | | | | | | M v | M | | Osz.-------*-----------------------------* | | ^ | | | | | *----------------------------*-------* I hope you can decifer my ASCIi-Painting... Seasons greetings, HolgerArticle: 2508
ecp@focus-systems.on.ca (Eric Pearson) writes: > Charles P. Ohrbom <chuck@aeroastro.com> wrote: > >I need to implement a UA(R)T, i.e. the transmitt portion only, in an Actel > >PLD. I will use a pin on a MC68332 as the UAR(T) receiver. Does anyone > >have a design they could pass along? > > > The 1991 version of the Actel handbook had a design for a UART in the > application notes. I seem to remeber the transmitter as being the > simpler of the two sections (just shift the data out with start/stop > bits added at the correct rate). > A bit off subject, since the poster wanted a transmitter: Yes the transmitter is easier, since the receiver needs a small state machine to detect the start bit and then start sampling in the middle of the bit. But the transmitter needs a bit counter, and there's an old trick to eliminate the counter for the receiver. Preset the receive shift register to ones. When the first zero (the start bit) gets shifted out, you're done. -- Ken Goldman kgold@watson.ibm.comArticle: 2509
In answer to your question about clock enables in Altera devices - Altera has a number of device primitives in their library which provide a clock enable. The most common of these is the "DFFE" primitive, which provides a D input, clock (rising-edge triggered), active-low preset and clear, and an active-high clock enable. There is also a "TFFE" ("T" flip-flop with enable), as well as JKFFE types. If you are using the MAX+PLUSII software, you can consult the on-line help for info on all the available primitives. One additional comment is in order: The 7000 series parts do indeed have clock enables as shown on the data sheets. If you use a primitive with clock enable in your design, and target the design for an Altera device which does not directly support clock-enables, the MAX+PLUSII compiler will automatically synthesize the logic necessary to provide the correct behavior. I can't answer #2, because I have not used the XNF writer before.Article: 2510
****** Most up-to-date on-line symposium program is available at ********* ****** http://www.cs.washington.edu/research/projects/lis/www/fpga96 ********* FPGA `96 Advance Program ------------------------ 1996 ACM/SIGDA Fourth International Symposium on Field-Programmable Gate Arrays February 11-13, 1996 Monterey Beach Hotel, Monterey, California, USA Sponsored by ACM SIGDA, and Xilinx, Inc., Altera Corp. and Actel Corp. Over the past ten years FPGAs have revolutionized the way many systems are designed by providing a low-cost, fast-turnaround implementation alternative. This is an exciting time in an exciting field that is still expanding as new technologies appear, new architectures are proposed, and new CAD tools are developed to address problems specific to FPGAs. This Symposium focuses on the architectural and algorithmic issues that FPGA architects and CAD designers face today and in the future. This is a forum where researchers from industry and university present and debate the latest ideas in FPGA design and application. The technical program consists of papers concerning both the practical and theoretical aspects of FPGA architecture, CAD algorithms for using and testing FPGAs, and applications. The Symposium will be of interest to those developing FPGA architectures, both at the chip and board level, and those developing CAD algorithms for FPGAs. The Symposium is not of direct interest to immediate users of FPGAs. General Chair: Jonathan Rose, University of Toronto Program Chair: Carl Ebeling, University of Washington Publicity Chair: Jason Cong, UCLA Local Chair: Pak Chan, UC Santa Cruz Finance Chair: Steve Trimberger, Xilinx Program Committee Michael Butts, Quickturn Pak K. Chan, UCSC Paul Chow, U. Toronto Jason Cong, UCLA Ewald Detjens, Mentor Carl Ebeling, U. Washington Gareth Jones, Pilkington Dwight Hill, Synopsys Brad Hutchings, BYU Sinan Kaptanoglu, Actel Jonathan Rose, U. Toronto Richard Rudell, Synopsys Rob Rutenbar, CMU Takayasu Sakurai, Toshiba Martine Schlag, UCSC Tim Southgate, Altera Steve Trimberger, Xilinx Nam-Sung Woo, ATT Program Sunday February 11, 1996 6:00pm Registration 7:00pm Welcoming Reception, Monterey Beach Hotel, Monterey Monday February 12, 1996 7:30am Continental Breakfast/Registration 8:20am Opening Remarks Session 1: Novel FPGA Architectures Chair: Jonathan Rose, University of Toronto 8:30am Hybrid FPGA Architecture, A. Kaviani and S. Brown, University of Toronto 8:50am Plasma: An FPGA for Million Gate Systems, V.R. Amerson, R. Carter, W. Culbertson, P. Kuekes, G. Snider, L. Albertson, HP Labs 9:10am Flexible FPGA Architecture Realized of General Purpose Sea of Gates, K. Azegami, S. Kashi- wakura, K. Yamashita, Fujitsu Laboratories Posters: Novel FPGA Architectures 9:30-10:30am Coffee & Posters Session 2: Logic Module Design Chair: Richard Rudell, Synopsys 10:30am Using BDDs to Design ULMs for FPGAs, Z. Zilic and Z.G. Vranesic, University of Toronto 10:50am Series-Parallel Functions and FPGA Logic Module Design, S. Thakur, D.F. Wong, University of Texas, Austin 11:10am Combined Spectral Techniques for Boolean Matching, E. Schubert, W. Rosenstiel, University of Tuebingen Posters: Logic Module Design 11:30-12:00 LUNCH 12:00 - 1:30 Session 3: Performance Issues Chair: Steve Trimberger, Xilinx 1:30pm The Wave Pipeline Effect on LUT-Based FPGA Architectures, E.I. Boemo, S. Lopez-Buedo, J.M. Meneses, Universidad Politecnica de Madrid 1:50pm Timing Optimization for Hierarchical Field- Programmable Gate Arrays, V.C. Chan, D.M. Lewis, University of Toronto 2:10pm Technology Mapping of Sequential Circuits for LUT-Based FPGAs for Performance, P. Pan, C.L. Liu, Clarkson University Posters: Performance Issues 2:30-3:30pm Coffee & Posters Session 4: Theoretical Issues in Routing Architectures Chair: Jason Cong, UCLA 3:30pm A Method for Generating Random Circuits and Its Application to Routability Measurement, J. Darnauer and W.W-M. Dai, University of California, Santa Cruz 3:50pm Entropy, Counting, and Programmable Interconnect, A. DeHon, MIT 4:10pm Universal Switch Modules for FPGA Design, Y-W. Chang, D.F. Wong, C.K. Wong, University of Texas, Austin Posters: Theoretical Issues in Routing Architectures 4:30-6:00pm Free time/Posters Dinner 6:00-7:30pm 7:30-9:00pm PANEL FPGAs vs. Gate Arrays and Processors: Who Will Win? The FPGA industry has enjoyed rapid growth in the past ten years in terms of chip density and speed as well as ASIC market share. In the same period, however, we have also observed significant advances in all sectors of the semi- conductor industry -- state-of-the-art gate arrays have a capacity of over 10 million transistors and enable the `system-on-a-chip'. Design automation tools have made semi-custom designs much faster and easier to achieve while yielding both high density and high performance. High-end microprocessors have reached over 250 Mhz and can satisfy the needs of many real-time control and DSP/multi-media applications. New rapid prototyping technologies, such as laser-programmed gate arrays, have emerged for high-speed high-density prototyping. Given such a dynamic industry undergoing exponential growth, it is interesting to ask where FPGAs will stand five or ten years from now in the wide spectrum of design technologies. Will its share of the ASIC market continue to increase, or will it become more of a niche technology? It is likely that the relative importance of these technologies will change drastically over the next five to ten years. This panel comprises technology experts in the competing areas of FPGAs, gate arrays, processors and other technologies. They will focus on the technological and economic issues that give one implementation medium an advantage over others and discuss how new technologies and architectural developments may change the competitive balance. They will discuss the past, present and future of the technological forces driving the industry and debate where those forces are likely to take us in the future. Tuesday February 13, 1996 Session 5a: Field-Programmable Analog Arrays Chair: Paul Chow, University of Toronto 8:30am Design and Implementation of a Field- Programmable Analogue Array, A. Bratt and I. Macbeth, Pilkington Microelectronics 8:50am The EPAC Architecture: An Expert Cell Approach to Field-Programmable Analog Arrays, H.W. Klein, IMP Posters: Field-Programmable Analog Arrays 9:10-9:40am Coffee & Posters Session 5b: Testing Chair: Martine Schlag, UC Santa Cruz 9:40am Diagnosing Programmable Interconnect Systems for FPGAs, D. Ashen and F. Lombardi, Texas A&M University 10:10am Evaluation of FPGA Resources for Built-In Self- Test of Programmable Logic Blocks, C. Stroud, P. Chen, S. Konala, M. Abramovici, University of Kentucky Posters: Testing 10:30-11:00am Coffee & Posters Session 6: The Future of Fuse and SRAM FPGA Technologies Chair: Tim Southgate, Altera 11:00am Two invited speakers will present the state of the art in (anti-)fuse and SRAM technologies and discuss the impact of recent developments in these technologies on future architectures. Posters: FPGA Vendors 11:40-12:00 LUNCH 12:00 - 1:30 Session 7: Applications Chair: Dwight Hill, Synopsys 1:30pm DPGA Utilization and Application, A. DeHon, MIT 1:50pm Integrating Software with Run-Time Re- configured Hardware, M.J. Wirthlin and B.L. Hutchings, Brigham Young University 2:10pm Computing the Discrete Fourier Transform on Virtual Systolic Arrays, C. Dick, La Trobe University Posters: Applications 2:30-3:30pm Coffee & Posters Session 8: Design Systems Chair: Pak Chan, UC Santa Cruz 3:30pm RASP: A General Logic Synthesis System for SRAM-based FPGAs, J. Cong, J. Peck, UCLA, and Eugene Ding, AT&T Bell Laboratories. 3:50pm Emerald - An Architecture-Driven Tool Compiler for FPGAs, D. Cronquist and L. McMurchie, University of Washington 4:10pm Structured Design Implementation - A Strategy for Implementing Regular Datapaths on FPGAs, A. Koch, Technical University, Braunschweig Posters: Design Systems 4:30-5:00 5:00pm Symposium Ends. Hotel Information ----------------- The Symposium will be held at the Monterey Beach Hotel, 2600 Sand Dunes Dr., Monterey, CA 93940, USA. The phone number for room reservations is 1-800-242-8627 or +1-408-394-3321 (Fax +1-408-393-1912). Reservations must be made before January 6, 1996. Identify yourself with the group Association for Computing Machinery FPGA `96 Symposium to receive the special Symposium rates, which are $75 for single or double Gardenview and $105 for single/double Oceanview. Parking is free. Check- in time 4pm. Directions to Hotel: From San Jose (a 1.5 hour trip) or San Francisco Airport (2.5 hrs) take HWY 101 South to HWY 156 West to HWY 1 South. On HWY 1 South, take Seaside/Del Rey Oaks exit. The hotel is at this exit, on the ocean side. You can also fly directly to the Monterey Airport, which is served by United, American and other airlines with at least 8 flights per day. FPGA `96 REGISTRATION --------------------- The Symposium registration fee includes a copy of the symposium proceedings, a reception on Sunday evening, February 11, coffee breaks, lunch on both days, and dinner Monday evening, February 12. First Name:___________________________________________ Last Name:____________________________________________ Company/Institution___________________________________ Address:______________________________________________ City:___________________State:________________________ Postal Code:_______________Country:____________________ Email:__________________________________________________ Phone:_______________________Fax:_______________________ ACM Member #____________ Circle Fee: Before January 25, 1996 After January 25, 1996 ACM/SIGDA Member US $320 US $390 *Non-Member US $420 US $490 Student US $90 US $90 (does not include reception or banquet, available for $20 and $35 respectively) *If you are not an ACM/SIGDA member we are giving you the opportunity to join by paying your first year's dues out of your conference non-member registration fee -- a US$100 value. Forms will be available at on-site registration. Guest Reception Tickets #Tickets______x US $20 ______ Guest Banquet Tickets #Tickets______x US $35 ______ Total Fees:____________________(Make checks payable to ACM/FPGA'96) Payment Form (Circle One): AMEX MASTERCARD VISA CHECK Credit Card#:____________________________________ Exp. Date:_______________________________________ Signature:_______________________________________ Send Registration with payment to: FPGA `96 - Colleen Matteis, 553 Monroe St., Santa Clara, CA. 95050, USA. Phone: +1(408)296-6883 Fax: +1(408)985-8274. For registration information contact Colleen Matteis, e-mail: sigda@nextwave.com, or cmatteis@aol.com. Cancellation must be in writing, and received by Colleen Matteis before January 24,1996.Article: 2511
Can anyone recommend some reading material to bring me up to speed on programmable logic devices. Specially useful would be any 3'rd party material on the subject using the Lattice Semiconductor's products. I am real new to programmable logic (abt 3 weeks). I'm using the Lattice ispLSI 1000 and 2000 families. Mainly so that I don't have to buy a programmer. Any recommendations on free or shareware software especially simulation software would be appreciated. Thanks. George..Article: 2512
I have a need for a "C" size VXI module that contains a couple of re-progarmable FPGA modules. Does anyone know of a device such as this. Module needs to have the VXI communication interface and the circuitry to accept the design personalazation data and program the local FPGA's. The majority of the FPGA I/O pins needs to commited to connectors on the front of the VXI module.Article: 2513
In article <Bb5Px*T-1@wolf359.exile.org> eric@wolf359.exile.org (Eric Edwards) writes: I am fishing for trends here... In an effort to kick start my engineering career into existance, I intend to learn one of these HDLs soon. The question is, which one? Industry seems not to hire generalists so it seems best to put the energy into one rather that divide among both. Which HDL would provide the greatest benefit for someone trying to break into logic design? You will get no agreement from me that industry does not hire generalists. With trends, as you mention, changing so often, generalists are oh so important. The choice of HDL is not the issue. Learn them both, they will both be around a long time. The real important thing to learn in logic design is what you are trying to create. View the HDL as the medium, and your goal is to get across a message. You want to create a easily verifiable, synthesizable description of hardware, which will run at high clock speeds while implementing complex protocols. Focusing on Verilog vesus VHDL is like an artist focusing on the decision between using watercolor and oil paints. What the artist should be considering is making delightful depictions of interesting scenes that catch the viewers interest. You should seek to set up in yourself the ability to be proficient in any and all the tools the design shop in which you want to work will be using. Bringing proficiency in some existing tools is useful. Better is to bring the ability to critically evaluate the tools in use, and the capability to create a masterpiece in whatever medium your might have at your disposal. Michael McNamara. -- Michael McNamara Verilog-HDL Consulting Services, Inc.Article: 2514
Stephen L. Wasson of HighGate Design, Inc. Saratoga, CA is my idea of a florplaning GURU. He lectures at Xilinx Advanced Seminars and at various conferences. He will be speaking at SuperCon 96. His papers on this subject are usually available by calling him and asking for them. I notice he has a new article in the Jan 96, Integrated System Design. He floor plans before designing the circuit so that his design will best take best advantage of that part of the chip where it must reside to be optimal. I would recommend that you read as many of Stephen's articles as you can scrounge, and then use the new floor planner in the 6.0 Xilinx tools to implement Stephen's philosophies. There is also an animated tutorial on floor planing that comes with the new Xilinx tools. Dave Decker Diablo Research Corp. ddecker@diablores.com ddecker@usa.pipeline.com -- mushArticle: 2515
In article <DI6H3x.5Lz@world.std.com>, John Cooley <jcooley@world.std.com> wrote: >Jyri Hamalainen <jyrih@cat.co.za> wrote: >>Does anyone here know anything about protecting ASIC's from reverse >>engineering? OR logic camourflaging? With modern technologies such as >>electron beam induced current imaging and Charge induced voltage >>alteration (Scania Labs), I suppose there are no more solutions to >>protecting ones investment? This isn't protecting one's investment, it's cheating the purchaser of ones products by making them unrepairable. Built in obsolescence is what this amounts to. >Jyri, one clever idea for protecting designs in chips I heard of was >using Xilinx FPGA's that were programmed once at the factory with a >small battery attached after programing. (That is, the power-up >program for the Xilinx part was NOT included in the PCB.) What this >did was make the circuit unreversable but still functional. Such designs should only be used for top-secret military type equipment, never for equipment sold for commercial or consumer use. Anyone who designs such equipment for commercial or cosumer use is a creep and an idiot. My reasoning is that after some number of years, someone who PAID for this equipment may wish to cotinue using it, and may want to try to repair it. -- R. D. Davis * http://www.access.digex.net/~rdd \Computer preservationist. Home: +1 410 744-7964 * Eccentrics have more fun! :-)\Unwanted systems gladly Unconventional Computer Consulting & PERQ Software, \disassembled, removed divs. of Transpower Industries, Inc. +1 410 744-4900 \for free and preserved.Article: 2516
To implement a clock enable in Verilog, do the following : always @ (posedge clk or posedge reset) begin if (reset) q <= 1'b0; // or q <= 1'b1; (for preset) else if (clken) q <= d; else q <= q; end The begin/end and the last else condition are redundant, but cleaner to read. -JoeArticle: 2517
In article <Bb5Px*T-1@wolf359.exile.org>, Eric Edwards <eric@wolf359.exile.org> wrote: >I am fishing for trends here... > >In an effort to kick start my engineering career into existance, I intend >to learn one of these HDLs soon. The question is, which one? Industry >seems not to hire generalists so it seems best to put the energy into one >rather that divide among both. Which HDL would provide the greatest benefit >for someone trying to break into logic design? Actually, I think they both are suitably similar to simply say that if you learn one, you've learned the other. On your resume, just put down, "VHDL/Verilog HDL". I've seen more job wanted ads that say "VHDL" than "Verilog HDL", so it looks like VHDL is a bit more popular. I could be wrong, though. Ciao for now, - Steve Weigand (weigand@marlin.ssnet.com)Article: 2518
> Such designs should only be used for top-secret military type > equipment, never for equipment sold for commercial or consumer use. > Anyone who designs such equipment for commercial or cosumer use is a > creep and an idiot. My reasoning is that after some number of years, > someone who PAID for this equipment may wish to cotinue using it, and > may want to try to repair it. Huh? I don't follow your argument at all. How does implementing some form of protection against reverse-engineering stops an owner from repairing the equipment which contains the protected chip? After all, ASICs are custom, single chips. Nobody is going to "repair" a bad ASIC. If you suspect you have a bad one the only thing you can do is replace the chip. To do that you need to go to the chip manufacturer to get a new one. How is this any different if the manufacturer tries to take steps to guard his chip against reverse engineering? C. ========================================================================== Charlie Hubbard | As the great philosopher Bingo once chubbard@oneworld.owt.com | said, "I have seen the future and http://www.owt.com/users/chubbard | Java is its name-o." ==========================================================================Article: 2519
Do what you may. It will only slow down the reverse engineering but never stop it. The only thing you can do it to keep putting out great products, obsoleting your own stuff every 2 years.Article: 2520
In <4beo90$b8c@access1.digex.net> rdd@access1.digex.net (R. D. Davis) writes: > >In article <DI6H3x.5Lz@world.std.com>, >John Cooley <jcooley@world.std.com> wrote: >>Jyri Hamalainen <jyrih@cat.co.za> wrote: >>>Does anyone here know anything about protecting ASIC's from reverse >>>engineering? OR logic camourflaging? With modern technologies such as >>>electron beam induced current imaging and Charge induced voltage >>>alteration (Scania Labs), I suppose there are no more solutions to >>>protecting ones investment? > >Anyone who designs such equipment for commercial or cosumer use is a >creep and an idiot. You may have not have noticed that the question referred to ASICS, which by their very nature, cannot easily be replaced by off the shelf stuff. They are custom parts. (application specific integraded circuits) Typically a lot of time and money goes into designing these parts.Article: 2521
Can anyone give me a quick synopsis of the number of each type of routing connection (single length, double length, longlines) present in each row and column of the 4025E device? I am trying to determine if a large datapath design will fit. The entire design will be manually placed, but has a very regular structure. The only information I have found so far is that there are 8 single length lines per row/column and 6 longlines. However I have come across a note that states that only some of the horizontal longlines have TBUFs. The design would require 4 such lines per row. I would like to verify this and obtain more specific information. Also, has anyone had any experience with implementing small multiplier arrays in the 4000E series devices? I know that it is possible, but the question is whether the resulting performance is good enough to be worth a larger device size. Thanks in advance for any information or stories. ---------------------------------------------------------------- Ewan D. Milne / Computervision Corporation (milne@petra.cv.com)Article: 2522
In article <4bpe8r$bd7@hettar.cv.com> milne@cv.com (Ewan D. Milne) writes: >Can anyone give me a quick synopsis of the number of each type >of routing connection (single length, double length, longlines) >present in each row and column of the 4025E device? I am trying >to determine if a large datapath design will fit. The entire >design will be manually placed, but has a very regular structure. > >The only information I have found so far is that there are 8 >single length lines per row/column and 6 longlines. However I >have come across a note that states that only some of the horizontal >longlines have TBUFs. The design would require 4 such lines per >row. I would like to verify this and obtain more specific information. > >Also, has anyone had any experience with implementing small >multiplier arrays in the 4000E series devices? I know that it >is possible, but the question is whether the resulting performance >is good enough to be worth a larger device size. > >Thanks in advance for any information or stories. > >---------------------------------------------------------------- >Ewan D. Milne / Computervision Corporation (milne@petra.cv.com) The 4025E has the same routing resources as the 4025, and on a per tile basis, it is the same as the 4003, 4003H, 4005, 4005H, 4006, 4008, 4010, 4010D, 4013, 4013D, and all of their 'e' series equivalents, where appropriate. All of these devices have 8 global clock nets, 8 locals per channel (horizontal and vertical), 4 doubles per channel ( H and V), and 6 long lines (H and V). Only 2 of the horizontal longlines per row have tbufs on them, and everything in the arcitecture is biassed to be used as 2 bits per row, with datapath elements being vertical structures, and the datapath bus running orthogonal to these structures (i.e. using the horizontal longlines for distributing bi-directional data). The E series has a few extra pips between the global lines and the CLBs, but not enough to have a serious effect. There are also some minor routing changes inside the CLB, plus of course the dual port sync RAM capability. E series is also available in speed grades that are significantly faster than what's available in XC4000 devices. If you really need 4 bidirectional bits per row, then this is not going to work for you. Although.... The tbuf lines can be split in the middle, allowing the lines on the left and right side of the chip to be independent. There are some cases where this can be quite useful. As always, the success of fast complex datapaths in the larger devices is very dependent on the quality of the floorplan. If you write about the design you are doing (in a lot more detail), then I could make some recomendations. I have built over 100 small multipliers in XC4000E in a bunch of designs. They were quite specific to an application, and could hardly be called 'general purpose'. They were all "multiply number (8 to 14 bits) by a constant", with some interesting constraints on what the constants could be. Some of these multipliers generated outputs 24 bits wide. All ran at about 33MHz in various 4006E-3, 4008E-3, and 4010E-3. All included a register on the output. Although not very 'macho', I prefer to stay away from the bragging-rights devices like the 4025(E), and do my designs in a pair of 4010(E) or 4013(E) devices. It is easier to route, the designs run faster, it forces me to do the partitioning, and it is cheaper. Hope some of this is of help :-) Philip FreidinArticle: 2523
R. D. Davis (rdd@access1.digex.net) wrote: : This isn't protecting one's investment, it's cheating the purchaser : of ones products by making them unrepairable. [snip] : My reasoning is that after some number of years, someone who PAID for : this equipment may wish to cotinue using it, and may want to try to : repair it. When someone buys something, they should expect proper operation for some time, but not forever. If it can be repaired, great, but it can provide value even if it fails later on. That's not cheating the purchaser. With electronic components/assemblies, the manufacturer has configured some materials for you, in order to perform some function you desire. You pay them for the product, and you get that function for at least the warranty period or some reasonable time for the type of device. So you make a protected chip, and make it to last some number of hours minimum. If it fails before then, it gets replaced free. If it lasts beyond its expected lifetime, that's free use of a product, which is a benefit to the purchaser. If a product is not used as intended and it fails, or if it's tampered with and self-destructs, the purchaser eats the loss. That's the way it's always been.Article: 2524
In article <4bqt7v$q99@earth.superlink.net>, rob-l@superlink.net (Rob-L) wrote: > With electronic components/assemblies, the manufacturer has configured > some materials for you, in order to perform some function you desire. > You pay them for the product, and you get that function for at least the > warranty period or some reasonable time for the type of device. > > So you make a protected chip, and make it to last some number of hours > minimum. If it fails before then, it gets replaced free. If it lasts > beyond its expected lifetime, that's free use of a product, which is a > benefit to the purchaser. If a product is not used as intended and it > fails, or if it's tampered with and self-destructs, the purchaser eats > the loss. That's the way it's always been. OK -- I just want to clarify one thing. If the reverse-engineered-protected chip includes a clock that registers the number of cycles that the chip has been used, at which point it shuts itself down, never to work again, then this chip is the ideal to which your organization aspires?? -- www/ftp directory: ftp://ftp.netcom.com/pub/hb/hbaker/home.html
Site Home Archive Home FAQ Home How to search the Archive How to Navigate the Archive
Compare FPGA features and resources
Threads starting:
Authors:A B C D E F G H I J K L M N O P Q R S T U V W X Y Z