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
_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_| _| _| _| _| _| _|_|_|_| _|_|_|_| _| _| _|_|_|_| _|_|_|_| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _|_|_| _|_|_|_| _| _|_|_|_| _|_|_|_| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _| _|_|_|_| _|_|_|_| _|_|_|_| _| _| _| _| _| _|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_| F P L '98 E I G H T H I N T E R N A T I O N A L W O R K S H O P FIELD PROGRAMMABLE LOGIC AND APPLICATIONS (http://xputers.informatik.uni-kl.de/FPL/index_fpl.html) Advance Program =============== _|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_| _| _| _| August 31 - September 3, 1998 (Monday - Thursday) _| _| _| _| Tallinn, Estonia _| _| _| _|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_|_| Aim: From tinkertoy to parallel computing paradigm ==== The methodology of reconfigurable circuits and systems is evolving from tinkertoy approach to an: Innovative Parallel Computing Paradigm which combines computing in time with computing in space. The aim of this workshop is to bring together workers from throughout the world for a wide ranging discussion of all forms of field programmable logic, particularly field programmable gate arrays and complex programmable logic devices, and their applications. It is intended to discuss the increasing range of device types, industrial applications, advanced design tool development, research applications, novel system architectures and educational experiences. The workshop will include regular presentations, posters and discussion sessions, and it is expected that most of the delegates will wish to make some contribution to one or more of these. The workshop is the eighth in a series of workshops which were held in Oxford (1991, 1993 and 1995), Vienna (1992), Prague (1994), Darmstadt (1996) and London (1997). The following program is subject to change without notification. ==================================================================== Sunday, Aug. 30 ==================================================================== -------------------------------------------------------------------- 20.00 Informal Gathering and Registration at the Hotel Viru (lobby) -------------------------------------------------------------------- ==================================================================== Monday, Aug. 31 ==================================================================== -------------------------------------------------------------------- 8.00 - 9.00 Registration at Tallinn National Library -------------------------------------------------------------------- 9.00 - 9.20 Opening Session -------------------------------------------------------------------- 9.20 - 10.20 Keynote -------------------------------------------------------------------- V.Milutinovic, University of Belgrade: Key Issues in Reconfigurable Computing -------------------------------------------------------------------- 10.20 - 10.35 Coffee Break -------------------------------------------------------------------- 10.35 - 12.15 Session 1: Design Methods -------------------------------------------------------------------- D.Robinson, P.Lysaght, G.McGregor, U Strathclyde: New CAD Framework Extends Simulation of Dynamically Reconfigurable Logic W.Luk, S.McKeever, IC London: Pebble: A Language For Parametrised and Reconfigurable Hardware Design V.Sklyarov, R.Sal Monteiro, N.Lau, A.Melo, A.Oliveira, K.Kondratjuk, Aveiro U: Integrated Development Environment for Logic Synthesis of Digital Circuits Based on Dynamically Reconfigurable FPGAs R.Hartenstein, M.Herz, F.Gilbert, U Kaiserslautern: Designing for the Xilinx XC6200 FPGAs -------------------------------------------------------------------- 12.15 - 13.30 Lunch -------------------------------------------------------------------- 13.30 - 15.10 Session 2: General Aspects -------------------------------------------------------------------- J.Becker, A.Kirschbaum, F.-M.Renner, M.Glesner, TU Darmstadt: Perspectives of Reconfigurable Computing in Research, Industry and Education G.Brebner, U Edinburgh: Field-Progammable Logic: Catalyst for New Computing Paradigms W.Luk, NShirazi, P.Y.K.Cheung, IC London: Run-time management of partially-reconfigurable designs M.Platzner, G.De Micheli, Stanford U: Acceleration of Satisfiability Algorithms by Reconfigurable Hardware -------------------------------------------------------------------- 15.10 - 15.30 Coffee Break -------------------------------------------------------------------- 15.30 - 17.10 Session 3: Prototyping / Simulation -------------------------------------------------------------------- J.Stohmann, K.Harbich, M.Olbrich, E.Barke, U Hannover: An Optimized Design Flow for Fast FPGA-Based Rapid Prototyping H.Krupnova, G.Saucier, INP Grenoble: A Knowledge-Based System for Prototyping on FPGAs R.Macketanz, W.Karl, TU Munich: JVX - A Rapid Prototyping System Based on Java and FPGAs J.Shetler, B.Hemme, C.Yang, C.Hinsz, Cal Poly: Prototyping New ILP Architectures Using FPGAs -------------------------------------------------------------------- 18.30 Reception at Tallinn Town Hall -------------------------------------------------------------------- ==================================================================== Tuesday, Sep. 1 ==================================================================== -------------------------------------------------------------------- 8.30 - 10.10 Session 4: Development Methods -------------------------------------------------------------------- Samary Baranov, Ben Gurion U Negev: CAD System for ASM and FSM Synthesis J.M.Emmert, A.Randhar, D.Bhatia, U Cincinnati: Fast Floorplanning for FPGAs M.Renovell, J.M.Portal, J.Figueras, Y.Zorian, LIRMM-UM2: SRAM-Based FPGAs: A Fault Model for the Configurable Logic Modules G.Haug, W.Rosenstiel, FZI Karlsruhe: Reconfigurable Hardware as Shared Resource in Multipurpose Computers -------------------------------------------------------------------- 10.10 - 10.35 Coffee Break -------------------------------------------------------------------- 10.35 - 11.50 Session 5: Accelerators -------------------------------------------------------------------- S.Robinson, M.Caffrey, M.Dunham, LANL Reconfigurable Computer Array: The Bridge Between High Speed Sensors and Low Speed Computing W.Luk, P.Andreou, N.Shirazi, D.Siganos, IC London: A Reconfigurable Engine for Real-Time Video Processing F.-M.Renner, J.Becker, M.Glesner, TU Darmstadt: An FPGA Implementation of a Magnetic Bearing Controller for Mechatronic Applications -------------------------------------------------------------------- 11.50 - 13.05 Session 6: System Architectures -------------------------------------------------------------------- R.Hartenstein, M.Herz, T.Hoffmann, U.Nageldinger, U Kaiserslautern: Exploiting contemporary memory techniques in reconfigurable accelerators A.Donlin, U Edinburgh: Self Modifying Circuitry - A Platform for Tractable Virtual Circuitry K. GajjalaPurna, K.Simha, D.Bhatia, U Cincinnati: REACT: Reactive Environment for Runtime Reconfiguration -------------------------------------------------------------------- 13.05 - 14.20 Lunch -------------------------------------------------------------------- 14.20 - 16.00 Session 7: Applications -------------------------------------------------------------------- S.Charlwood, P.James-Roxby, U Birmingham: Evaluation of the XC6200-series architecture for cryptographic applications A.Zakerolhosseini, P.Lee, E.Horne, U Kent: An FPGA based object recognition machine G.Acher, W.Karl, M.Leberecht, TU Munich: PCI-SCI Protocol Translations: Applying Microprogramming Concepts to FPGAs T.Callahan, J.Wawrzynek, UC Berkeley Instruction-Level Parallelism for Reconfigurable Computing -------------------------------------------------------------------- 14.20 - 15.35 Session 8: Poster Introduction (parallel to S.7) -------------------------------------------------------------------- 15.35 - 16.20 Poster Exhibition (partially par. to S.7) -------------------------------------------------------------------- N.L.Miller, S.F.Quigley, U Birmingham: A Novel Field Programmable Gate Array Architecture for high speed arithmetic processing D.MacVicar, S.Singh, XILINX: Accelerating DTP with Reconfigurable Computing Engines C.N.Ojeda-Guerra, R.Esper-Chain, M.Estupinan, A.Suarez, U Las Palmas: Hardware Mapping of a Parallel Algorithm for Matrix-Vector Multiplication G.Brebner, U Edinburgh: An Interactive Datasheet for the Xilinx XC6200 N.Woolfries, P.Lysaght, S.Marshall, G.McGregor, D.Robinson, U Strathclyde: Fast Adaptive Image Processing in FPGAs using Stack Filters S.Sawitzki, A.Gratz, R.Spallek, U Dresden: Increasing Microprocessor Performance with Tightly-Coupled Reconfigurable Logic Arrays N.Bergmann, P.Sutton, Queensland U: A High-Performance Computing Module for a Low Earth Orbit Satellite using Reconfigurable Logic S.Yamagiwa, M.Ono, T.Yamazaki, P.Kulkasem, M.Hirota, K.Wada, U Tsukuba: Maestro-Link: A High Performance Interconnect for PC Cluster T.Shiozawa, K.Oguri, K.Nagami, H.Ito, R.Konishi, N.Imlig, NTT: A Hardware Implementation of Constraint Satisfaction Problem Based on New Reconfigurable LSI Architecture P.Merino, J.Lopez, M.Jacome, U Politecnica de Madrid: A Hardware Operating System for Dynamic Reconfiguration of FPGAs E.Cerro-Prada, P.B.James-Roxby, U Birmingham: High speed low level image processing on FPGAs using distributed arithmetic T.-T.Do, H.Kropp, C.Reuter, P.Pirsch, U Hannover: A Flexible Implementation of High-Performance FIR Filters on Xilinx FPGAs I.Vassanyi, U Veszprem: Implementing processor arrays on FPGAs S.Holmstrom, K.Sere, Abo Akademi U: Reconfigurable Hardware - A Study in Codesign C.Ackad, TU Braunschweig: Statechart-based HW/SW-Codesign of a Multi-FPGA-Board and a Microprocessor -------------------------------------------------------------------- 16.20 - 16.40 Coffee Break -------------------------------------------------------------------- 16.40 - 18.00 Session 9: Hardware/Software Codesign -------------------------------------------------------------------- G.McGregor, D.Robinson, P.Lysaght, U Strathclyde: A Hardware/Software Co-design Environment for Reconfigurable Logic Systems K.Bondalapati, V.Prasanna, U of Southern California: Mapping Loops onto Reconfigurable Architectures S.Asaad, K.Warren, IBM T.J.Watson Research Center: Speed Optimization of the ALR circuit using an FPGA with embedded RAM: A Design Experience -------------------------------------------------------------------- 19.30 Banquet at von Glehn's castle -------------------------------------------------------------------- ==================================================================== Wednesday, Sep. 2 ==================================================================== -------------------------------------------------------------------- 8.30 - 9.45 Session 10: System Development -------------------------------------------------------------------- R.Kress, A.Pyttel, A.Sedlmeier, Siemens AG: High-level Synthesis for Dynamically Reconfigurable HW/SW-Systems N.McKay, S.Singh, XILINX: Dynamic Specialisation of XC6200 FPGAs by Partial Evaluation S.Guccione, XILINX: Webscope: A Circuit Debug Tool -------------------------------------------------------------------- 9.45 - 10.00 Coffee Break -------------------------------------------------------------------- 10.00 - 11.40 Session 11: Algorithms on FPGAs -------------------------------------------------------------------- D.Lavenier, Y.Saouter, IRISA-CNRS: Computing Goldbach partitions using pseudo-random bit generator operators on a FPGA systolic array P.Zhong, M.Martonosi, S.Malik, P.Ashar, Princeton U: Solving Boolean Satisfiability with Dynamic Hardware Configurations J.Põldre, M.Mandre, K.Tammemäe, Tallinn Technical U: Modular exponentiator realization on FPGA B.Feher, G.Szedo, TU Budapest: Cost effective 2x2 inner product processors -------------------------------------------------------------------- 10.00 - 11.15 Session 12: Poster Introduction (parallel to S.11) -------------------------------------------------------------------- 11.15 - 12.00 Poster Exhibition (partially par. to S.11) -------------------------------------------------------------------- A.Touhafi, W.F.Brissinck, E.F.Dirkx, U Brussel: Simulation of ATM switches using Dynamically Reconfigurable FPGA's T.Rissa, T.Mäkeläinen, J.Siirtola, J.Niittylahti, Tampere U: Fast Prototyping Using System Emulators A.Dandalis, V.Prasanna, U of Southern California: Space-efficient Mapping of 2D-DCT onto Dynamically Configurable Coarse-Grained Architectures I.Lemberski, M.Ratniece, Riga Aviation U: XILINX4000 Architecture-Driven Synthesis for Speed V.Tomachev, Inst. of Eng. Cybernetics Belarus: The PAL-implementation of Boolean function characterized by minimum delay A.Abo Shosha, P.Reinhart, F.Rongen, Research Centre Jülich: Reconfigurable PCI-Bus Interface (RPCI) A.Trost, A.Zemva, B.Zajc, U Ljubljana: Programmable Prototyping System for Image Processing L.Palomar, T.Fukuda, E.Dadios, Nagoya U: A VHDL Design of an Arithmetic Co-Processor to Implement the Kohonen's Self-Organizing (Feature) Map J.Fischer, C.Müller, H.Kurz, TU Aachen: A Co-Simulation Concept for an Efficient Analysis of Complex Logic Designs A.Döring, W.Obelöer, G.Lustig, Med. U Lübeck: Programming and Implementation of Reconfigurable Routers M.Moure, U Vigo: Virtual Instruments based on reconfigurable logic C.Siemers, D.Möller, FH Westküste: The >S<puter: Introducing a Novel Concept for Dispatching Instructions Using Reconfigurable Hardware L.Lagadec, B.Pottier, U de Bretagne Occidentale: A 6200 model and editor based on object technology M.Eisenring, J.Teich, ETH Zuerich: Interfacing Hardware and Software J. Hwang, E. Dellinger, S. Mitra, S. Mohan, C. Patterson, R. Wittig, XILINX: Generating Layouts for Self-Implementing Modules -------------------------------------------------------------------- 12.00 - 13.00 Closing Session -------------------------------------------------------------------- T.Maruyama, T.Funatsu, T.Hoshino, U of Tsukuba: A Field-Programmable Gate-Array System for Evolutionary Computation T.Miyazaki, K.Shirakawa, M.Katayama, T.Murooka, A.Takahara, NTT: A Transmutable Telecom System Conclusions, Award Ceremony and Announcements -------------------------------------------------------------------- 14.00 Tour to Lahemaa National Park, visiting manors, and dinner in Altja tavern -------------------------------------------------------------------- ==================================================================== Thursday, Sep. 3 ==================================================================== -------------------------------------------------------------------- 8.30 - 12.00 Tutorial at Tallinn Tech. University -------------------------------------------------------------------- B. Radunovic, V. Milutinovic, Univ. Belgrade: A Survey of Reconfigurable Computing Architectures, Part 1 -------------------------------------------------------------------- 13.00 - 16.00 Tutorial at Tallinn Tech. University -------------------------------------------------------------------- B. Radunovic, V. Milutinovic, Univ. Belgrade: A Survey of Reconfigurable Computing Architectures, Part 2 -------------------------------------------------------------------- ==================================================================== Conference Venue: ----------------- Tallinn National Library, Tõnismägi 2, EE0100 Tallinn, Estonia Program Committee: ------------------ Peter Athanas, Virginia Tech, USA Samary Baranov, Ben Gurion U. Negev,Israel Stephen Brown, U. of Toronto, CA Klaus Buchenrieder, Siemens AG, FRG Steven Casselman, VCC, USA Bernard Courtois, INPG, Grenoble, France Carl Ebeling, U. of Washington, USA Norbert Fristacky, Slovak Technical U., SK Manfred Glesner, TH Darmstadt, FRG John Gray, Xilinx, UK Herbert Grünbacher, Vienna U., Austria Reiner Hartenstein, U. of Kaiserslautern, FRG Brad Hutchings, Brigham Young U., UAS Udo Kebschull, U. of Tübingen, FRG Andres Keevallik, Tallinn Technical U., Estonia Wayne Luk, Imperial College, UK Patrick Lysaght, U. of Strathclyde, Scotland Toshiaki Miyazaki, NTT Laboratories, Japan Will Moore, Oxford U., UK Wolfgang Nebel, U. of Oldenburg, FRG Paolo Prinetto, Politecnico di Torino, Italy Jonathan Rose, U. of Toronto, Canada Zoran Salcic, U. of Auckland, New Zealand Marc Shand, Digital Systems Research Center, USA Stephen Smith, Altera, USA Steve Trimberger, Xilinx, USA General Chairman: ----------------- Prof. Andres Keevallik Tallinn Technical University Raja 15 Tallinn EE-0026 Estonia Phone: +372-6104440 Fax: +372-6202246 email: akeev@cc.ttu.ee http://www.ttu.ee/fpl98 Program Chairman: ----------------- Prof. Reiner W. Hartenstein University of Kaiserslautern P. O. Box 3049 D-67653 Kaiserslautern Germany Phone: +49 631 205-2606 Fax: +49 631 205-2640 email: hartenst@rhrk.uni-kl.de http://xputers.informatik.uni-kl.de/FPL/index_fpl.htmlArticle: 10901
I see that some people have given you satisfactory answers to this question already. If you would like to save some work and use a ready-made function, I have provided a master function - fully tested and documented - written in Altera AHDL and which I provide free of charge. You can find this at http://www.acte.no/freecore - and follow the links to the "modules" section. Regards, Rune Baeverrud Bryn Wolfe wrote: > Has anybody tried to implement an I squared C interface on an FPGA? I am > trying to implement it in an HDL, say VHDL, but find that my logic is > excessive for such a simple serial interface. I figure I'm missing some > key insight into making the logic much simpler. > > If anybody has attempted this or knows where I might look for > information on this subject, please help. > > -- > Bryn Wolfe - Robotics Engineer > Metrica TRAC LabsArticle: 10902
Three suggestions: - Gzip sounds a bit heavyweight for your need, but it does compress very well. The space saved by better compression might offset the space needed for decompression code. Most of the run-time memory needed can go in the same place as the decompressed file. (Result of sliding window technique -- a little additional memory needed for huffman decoder though). The Gzip decompressor is quite feasible to code in assembler, if you use a pure sliding window and especially if you use fixed huffman code weights. Take a look at inflate.c in the Gzip or libz distributions. You can ignore all the error checking and buffering and so forth -- the sliding window expander/huffman decoder are pretty short when you get down to the minimum required. - Forget Bzip, it uses tons of memory and is slow. It's very good compression (better than Gzip), but inappropriate for this problem. - Sliding window like Gzip but without the Huffman coding -- just use a fixed no. of bits for offset and a length, perhaps with a few special cases. This is very similar to run-length encoding, but has the big advantage that it is good at representing repeated patterns with occasional differences interleaved. Sounds ideal for an FPGA bitstream, but I haven't tried it. You can modify Gzip to generate this data (by forcing fixed Huffman weights and changing the maximum offset). Like run-length encoding, the code to decompress this is simple, tiny and very fast. And it doesn't need any extra RAM. - Take a look on the net for the LZO decompressors, specifically designed for fast, simple decompression in embedded & multimedia systems. (LZO = Lempel-Ziv-Oppenheimer I think). The code includes several different compression schemes, compressors and decompressors for each. Some of the decompressors are available in i386 assembly language as well as C -- if that's not your processor, the code may still give ideas on coding for your CPU. As I said, these schemes are designed for fast & simple decompression. I wouldn't be surprised if some of the schemes do just as well as Gzip on an FPGA bitstream. I don't think they need extra RAM to run. Announcements for the code crop up on comp.os.linux.announce from time to time, so maybe Dejanews will help you find it. (It's not Linux specific, I just happen to have seen announcements there). Hope this helps, -- JamieArticle: 10903
Rickman wrote: > > The best way to deal with such things is for everyone to reply to the > sender. I have already sent him a message indicating what I think. If > enough people did that, it would make the posting not worth the junk > replies. In fact this guy took the time to send me a reply! > > The real problem is with the spammers that are advertising a web site or > post a long distance phone number. There is not much point in wasting > your time and money to reply to those. > > But if they post a valid email address, then they are fair game! > > Terry L. Zumwalt wrote: > > > > Dear Jeremy, > > > > I have toagree with the other newsgroup members. Your solicitation is not really > > what this news group is about. I my not sound as emotional as some of the other > > responses but my feelings are in agreement. Aside from the fact that I think > > your pyramid scheme preys on people who are desperate and need better help than > > you are offering. Please respect the wishes of the newsgroup and leave your > > advertisement off this thread. I would be surprised if you get even one positive > > response on this thread. > > > > Hope to never hear from you again. > > Best Regards > > Terry I find it's useful to CC the reply to postmaster@hisdomain, and root@hisdomain. If enough sysadmins get pissed off, maybe they will put a halt to this sh*t. -- Chris Fischer cfischer@frontiernet.net Owner, Coda Software, LimitedArticle: 10904
Goran.Bilski@enator.se wrote: > > I'm a senior consultant from Sweden who attended to the 35th DAC. I have ...snip... > Before DAC, I had seen a press release that Altera was going to show the > software that will come after MaxPlus2, and since I have been using MaxPlus2, > I was interested in what they had improved. So I walked to their demo suite > and stepped in and asked if I could see a demo. A man came towards me and and > said 'NO It is only for special invited people' and then turned his back on > me. I could do nothing but walk away. I'm still stunned about this attitude > to an unknown customer; he didn't care to find out who I was or what I was > doing. I might have been a customer who was going to buy thousands of their > FPGAs, but they didn't care. If they had just asked who I was, and then said > it is only for beta-site people and then asked me if I would like to become a > beta-site, that would have been o.k. For crying out loud, he could have > talked to me. If they had a good reason for not showing me the demo, I would > have accepted it. But no, they didn't care. Then I thought it might be only > this one guy with an attitude problem, so I walked to their booth again and > asked the same question, and I was given the same answer and attitude. > > So, if Altera is ignoring me I will start to ignore Altera. My company is ...snip... > You may say that I'm acting like a 5-year old kid, and that may be true. I > will ban Altera as long as I'm so disgusted with their attitude. I seldom get > angry, but when I do I get real angry. > > Since I'm a consultant I will use Altera only if my client says so, and if the > design really fits Altera. > Otherwise I will look for alternative architectures. > > Göran Bilski I feel your pain and I understand your response. There is no excuse for the way you were treated. No salesman should ever be so rude to anyone, even after finding out that the potential customer was a small dollar account. As engineers we have very little control over the products and sales approaches from the semiconductor companies. With the one exception: the purchase order. This is the sole reason for the existance of the sales force... sales. But then the PO is a powerful tool that should be used very carefully. Just like a gun, it can do harm to the person using it if not used carefully. Before you get Altera blacklisted with your company, you might try approaching your local sales office to see if they are interested in attempting to make up for the way you were treated. I have found that when you reach the people who are responsible for making you happy, they often will bend over backward to make up for some problem that was caused by someone else. Salesmen really are people too, honest! I have no connection with Altera. I don't even use their parts. So I have no axe to grind either way. In fact, if you are serious about sharing your VHDL designs, I would be very interested in the FIFO, the SHARC DSP link receiver/transmitter and the synthesizable conversion between IEEE Float and Std_Logic, as well as the behavior model for the SHARC DSP. I don't have an immediate need for these, but I can see use for them in Xilinx Virtex parts in the future. Thanks in advance. If you want to email me, please don't use the reply address. Use the address below. -- Rick Collins rickman@XYwriteme.com remove the XY to email me.Article: 10905
Simple vs. Complicated for an I2C interface is all a matter of implementation. If you want a fully compliant and robust interface, complete with bus arbitration, you will use much more than 20 CLB's. --- Tom Curran Memec Design Services -- Boston url: www.memecdesign.com email: tom_curran@memecdesign_dot_com phone: 978.266.9193 fax: 978.266.9194 vm: 800.688.7718 x3603Article: 10906
Simple vs. Complicated for an I2C interface is all a matter of implementation. If you want a fully compliant and robust interface, complete with bus arbitration, you will use much more than 20 CLB's. --- Tom Curran Memec Design Services -- Boston url: www.memecdesign.com email: tom_curran@memecdesign_dot_com phone: 978.266.9193 fax: 978.266.9194 vm: 800.688.7718 x3603Article: 10907
does it run at room temp with a slightly slowed clock? will it run at room temp with an elevated supply voltage? rk Philip Freidin wrote: > Given that your device runs when cold (internal paths and logic run > faster), but not at room temp, here are SOME options: > > CAE/Vendor issues: > 1) The speed files are not correct > 2) Back annotating speed info did not work correctly > 3) The simulator isn't using the speed info correctly > 4) The device is faulty > 5) Your design is relying on a path that the speed file doesn't cover > > Your issue: > 6) Asynchronous logic > 7) Paths that depend on N cycles taking N-1 cycles and logic > can't handle occasional metastables > 8) Timespec coverage is not complete > 9) Multiple clock domains, using more than 1 clock signal rather than CE > 10) External issue that depends on higher drive when device is cold. > > Let me write a little about (8) > > At the end of the trace report, it should indicate the percentage > coverage of your time specs. It should be over 99%. (for some designs it > is impossible to get 100% because of bugs in the way this coverage > percentage is calculated). > > Your design should not rely on any default specs. > > you can run trace from the command line with the following command: > > ¦trce mychip.ncd mychip.pcf -v 100 -u -o mychip.twr >> mychip.log > > The -v says give verbose info on the worst 100 paths for each timespec. > The report can be quite long. > > The -u says list unconstrained paths. These will be listed in the .twr > file in a section titled: > > ========================================================= > Timing constraint: Unconstrained path analysis > 293 items analyzed, 0 timing errors detected. > Minimum period is 13.289ns. > Maximum delay is 19.354ns. > --------------------------------------------------------- > > yada yada yada > > This section will be ommitted if ALL paths are covered by either specs you > entered, or default specs. Following this heading it will list (verbosely) > the worst 100 unconstrained paths. Go find them, and add tspecs to cover > them. Re-netlist, PAR, trce. Repeat till this section disappears, or the > paths are not constrainable (due to bugs in the SW. There are a few that > affect this section, but they are REALLY rare.) > > At the end of the .twr file, you need to see something like: > > Timing summary: > --------------- > > Timing errors: 0 Score: 0 > > Constraints cover 40047 paths, 0 nets, and 22452 connections (100.0% > coverage) > > Good luck. > Philip Freidin > > In article <35969ebb.5128088@news.dial.pipex.com> ems@see_sig.com (ems) writes: > > > >the silicon works fine when frozen. it fails after a couple of > >minutes at room temperature (20C, no cooling, to be precise), at 5V > >+/- 2%, on one tested device. > > > >note that i'm *not* (yet) claiming that there's a problem with the > >simulator or the speed file. however, the failure is very specific, > >and i'm sure, as far as i can be, that the simulator can't reproduce > >it. > > > >evan (ems@nospam.riverside-machines.com) > >Article: 10908
Hello, Connectix - the market leader in the design and manufacture of tethered video cameras- is looking to hire a Sr. Hardware Engineer. Experience: At least 5 years of board level design experience with FPGA prototyping is required as is knowledge of embedded microcontrollers and PCB layout techniques. Working knowledge of an HDL/ASIC design process is also required, in addition to an understanding of CCD and/or CMOS sensor technology. Contact: Jerel Cain Connectix Corporation http://www.connectix.com 650-295-7257 (phone) 650-571-8541 (fax) jcain@connectix.comArticle: 10909
Steve schrieb in Nachricht <3596c0ed.1378542@news.netspace.net.au>... >Hi > >We are currently running an embedded system with segmented Flash. >This allows us to have a Boot code section and a Main code section. ...snip... >And it brings the 40k file down to about 5k. But I would like more >compression! ...snip... >Anyone got any ideas? > >Steve > Keep in mind: As your XILINX Projects grows, your compression rate will go down !! It is easy to compress an EXO-file of an almost empty device. ManfredArticle: 10910
Hi, I know that Boundry Scan will let me test the FPGA at the inputs and outputs, but I would also like to be able to test internal paths. What are the tools and methods to test the internal paths of a PLD? Robert Peach Lockheed Martin Aeronautical Systems Atlanta, GA (770) 494-7587 rpeach@gelac.mar.lmco.comArticle: 10911
Hi again, I want to add to my last question. I would like to test the FPGA part so that it demonstrates proper operation under all combinations and permutations at the gates and pins of the device. I know that for very dense FPGAs that this is not possible as it would take to much time. I would like feedback on testing tools and methods that test an FPGA to a degree that I can be confident that I will not get a failure such as Latchup or SEU. Nothing is 100%, but if you know some things that are close I would appreciate hearing about it. Thank you. Robert Peach Lockheed Martin Aeronautical Systems Atlanta, GA (770) 494-7587 rpeach@gelac.mar.lmco.comArticle: 10912
Hi! I'm currently developing a very simple test board for the spartan XCS05 (PLCC84 package). I connected program (pin55), din (pin71) and cclk (pin73) using a 74ls14 schmidt-triger with a prallelport connector. Are these three signals all I need to programm the fpga in slave mode (I'm a little confudes with the datasheed)? Wich kind of output file should I generate (MCS, hex....) with F1.4 to download the bitstream (I want to write a litle transfer-software on my own)??? thanks in advance Marcus LankenauArticle: 10913
Robert Peach wrote: > > Hi again, > > I want to add to my last question. I would like to test the FPGA part > so that it demonstrates proper operation under all combinations and > permutations at the gates and pins of the device. I know that for very > dense FPGAs that this is not possible as it would take to much time. I > would like feedback on testing tools and methods that test an FPGA to a > degree that I can be confident that I will not get a failure such as > Latchup or SEU. Nothing is 100%, but if you know some things that are > close I would appreciate hearing about it. > Thank you. > > Robert Peach > Lockheed Martin Aeronautical Systems > Atlanta, GA > (770) 494-7587 > rpeach@gelac.mar.lmco.com Depends on what you mean. By the context, I think you are saying you want to do device testing independent of the design loaded in by the user? If this is the case, the degree of testing depends on the FPGA type. For an SRAM device, you can run a set of tests that will check the device fairly thoroughly. The manufacturer has some hooks in there that permit him to do more complete testing than the user however. If it is a one time programmable, then you as a user are more or less out of luck. Xilinx has an app note that breifly describes the reliability and some of the production testing. For more detail, I would suggest you contact you vendor of choice. The vendors have already performed a fairly exhaustive set of tests, so under an NDA you might be able to use their results to satisfy your needs or to set up additional testing. If you are talking about testing your design, then it obviously depends on the design. You can make your design more resistant to SEU by incorporating fault detection and correction in your logic. Controllability and observability of internal nodes will reduce the number of test vectors you need to get acceptable coverage of the design. You can get some access through the readback pins on some devices, but it is somewhat awkward to use. If you have the space and pins, add test access to the design. -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randraka The Andraka Consulting Group is a digital hardware design firm specializing in high performance FPGA designs for digital signal processing, computing and control applications.Article: 10914
ARM Ltd has developed a PCI board comprising of an ARM based Micro-controller and two Xilinx FPGAs which are available for user applications. It is possible that there might be a spare place on the beta trial which will take place over the next three months. If you are interested please read the extra information below and email me as soon as possible. Thanks Dominic Plunkett OptionExist Ltd Cambridge ASPIRE Beta Trial ================== The ASPIRE beta trial aims to test and evaluate the ASPIRE board and software packages that go with the board. The trial will run for three months, at the end we will require a report. OptionExist Ltd have been appointed by ARM Ltd to run the trial. If you have a use for the board and kit beyond the 3 month trial, it will be available at a discounted price. Please could you give us an indication as to the likely applications and which software packages you would require for the trial, and your experience with similar products. The ASPIRE Board ---------------- This ASPIRE development board was jointly developed between ARM Ltd and Oxford University under the ASPIRE project. The board contains the following: MicroController ( based on the ARM7TDMI). Clocked at 16MHz. BOOTROM with the Angel Debug Monitor. JTAG ICE many also be used 1/2Mbyte of Flash Memory for the MicroController. RS232 port DRAM SIMM (4/16Mbytes) System Controller (in an XC4013XL) which provides a DRAM controller, DMA Controller, Interrupt controller and other glue logic functions AMCC PCI chip which can directly DMA into/out of the DRAM via the system controller Selectable clock rates 4/8/16MHz. On board Crystal or PCI Clock. 2 x XC4013XL FPGA's for user Applications, programmed by the AT91 256K SSRAM directly connected to the FPGAs External power connector for standalone applications Many Expansion connectors Software -------- For the Trial we have arranged the following software to be available for evaluation if you need it. ARM Development Tool Version 2.11a (www.arm.com) HandelC : Oxford University Hardware Compilation Group have developed a C like language that can target the FPGA's. So it becomes very easy to develop hardware in a software environment. Oxford University now have a spin off company Embedded Solutions Ltd (www.embedded-solutions.ltd.uk). Xilinx Alliance standard, needed for HandelC or possible Xilinx Foundation which is needed for VHDL or HandelC. This software places and route the designs for the FPGA's. (www.xilinx.com) Reports -------- We are looking for people who could make use of the board in the 3 months. We would like monthly email reports about the board and the software. At the end of the trial we will require a final report, this should cover: What the board has been used for? How easy was it to use? Was the board suitable? Do you have enough information? Any problems you have had? It is useful to develop a product or test software? Do you know of any competing product? What improvements would you like? Dominic Plunkett OptionExist -----== Posted via Deja News, The Leader in Internet Discussion ==----- http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member ForumArticle: 10915
Computer Science and Engineering Department University of Nebraska - Lincoln Visiting Faculty / Post-Doctoral Fellow Positions Computer Science and Engineering Department University of Nebraska - Lincoln Applications are invited for Visiting Faculty / Post-Doctoral Fellow positions for one year to begin on August 17, 1998. Duties will include undergraduate and/or graduate instruction and research. Outstanding candidates with teaching and/or research experience in the following areas are sought: * VLSI design, design automation, and systems architecture, * Database and information systems, * Software engineering and human-computer interaction, * Communications, networking, and distributed systems, and * Theory and algorithms. Applicants at all levels of experience and with a Ph.D. in Computer Science, Computer Engineering, or a closely related field are eligible. Salary is negotiable and will be commensurate with qualifications and experience. These positions are non-tenure leading. However, the CSE Department has tenure-line vacancies and will be conducting searches during the coming academic year. Applications by visiting faculty for tenure-line positions will receive full consideration. The CSE Department currently has 20 full-time faculty lines (with three vacancies). The Department offers degrees programs both in computer science and in computer engineering leading to BS, MS, and PhD degrees. The CSE Department doctoral program recently was rated by the National Research Council second highest in the Big 8 Conference. The faculty research interests cover a wide range of topics in the defined focus areas of communication theory & foundations, distributed systems, human-centered software design, and visual information processing. Currently 100 students are enrolled in the graduate program and 500 in the undergraduate program. CSE Department faculty members are active participants in state-supported research centers, including the Center for Communication and Information Sciences, the Center for Technology Management and Information Science, the Center for Advanced Land Management Information Technologies, the Center for Biotechnology, and the Center for Infrastructure Research. Current funding for research from external sources is over $2.7M. Funding for active projects from the past year exceeded $800K with over $250K institutional support for more than 30 projects. Further information about the Department may be obtained from the web page at "http://cse.unl.edu/". As the largest educational institution in the state, UNL has since its founding in 1869 been a distinguished land grant institution. A member of the American Association of Universities, UNL also is a Carnegie I Research University and a member of the National Association of State Universities and Land Grant Colleges. There are nine undergraduate colleges and a law college as well as a graduate division granting degrees at the baccalaureate, master's and doctoral levels. Approximately 24,000 students matriculate annually. There are 32 doctoral programs, 63 master's programs, 130 undergraduate sequences and 14 pre-professional areas of study. The colleges are: Agricultural Sciences & Natural Resources, Architecture, Arts & Sciences, Business Administration, Engineering & Technology, Fine & Performing Arts, Human Resources and Family Sciences, Journalism & Mass Communications, Law, and Teachers. Screening begins immediately and will continue until suitable candidates are hired. Applicants should submit a letter of interest, a current CV and arrange to have three professional references sent as soon as possible, to: CSE Search Committee Chair (c/o: Ms. Linda Endres) University of Nebraska - Lincoln Lincoln, Nebraska 68588-0115 Telephone: (402)-472-5001 FAX: (402)-472-7767 E-mail: endres@cse.unl.edu WWW Site: http://cse.unl.edu The University of Nebraska is committed to a pluralistic campus community through Affirmative Action and Equal Opportunity and is responsive to the needs of dual career couples. We assure reasonable accommodation under the Americans with Disabilities Act. -- Department of Computer Science Stephen D. Scott Washington University, Campus Box 1045 sds@cs.wustl.edu One Brookings Drive http://www.cs.wustl.edu/~sds/ St. Louis, MO 63130-4899 (314) 935-4425Article: 10916
Does anybody have the Xilinx jedec fuse format for 9500 devices? TIA, Dennis Hawkins n4mwd@dont.spam.me.flinet.com (Remove "dont.spam.me." from address before replying) Do you want to know who has been calling you and hanging up when you answer the phone? Visit http://www.afn.org/~afn40110/telescum.htm to find out.Article: 10917
Check out our updated web-site at http://www.macrotechsemi.com. We offer Zero NRE ASICs. We can integrate RAM, ROM, CPU etc for designs up to 1M gates. Our process is 0.35 Micron 3.3V 5V tolerant. We will be updating soon with a detailed ASIC price estimator---look for that soon. Regards, KashArticle: 10918
Check out our updated web-site at http://www.macrotechsemi.com. We offer Zero NRE ASICs. We can integrate RAM, ROM, CPU etc for designs up to 1M gates. Our process is 0.35 Micron 3.3V 5V tolerant. We will be updating soon with a detailed ASIC price estimator---look for that soon. Regards, KashArticle: 10919
tom curran <tom_curran@memecdesign_dot_com> wrote in article <3597dfd3.593512078@mdsXboston>... > Simple vs. Complicated for an I2C interface is all a matter of > implementation. If you want a fully compliant and robust interface, > complete with bus arbitration, you will use much more than 20 CLB's. We did it in 24 CLBs in a 3142A and it is fully compliant (what ever that means...but it works with Dallas, National and Philips peripherals) and has arbitration for multiple masters (and that works too). So, it does depend on how you implement it ;-) Austin Franklin darkroom@ix.netcom.comArticle: 10920
If you want a fully compliant I2C interface you can't do it in a regular 5V FPGA at all. The I2C spec says that SDA and SCL can not be loaded when the power is removed from the chip. The parasitic diode due to the P channel pullup transistor in regular FPGA outputs makes this impossible. Perhaps you could use a 3V FPGA with 5V tolerant I/Os. This is not a problem if you only have one power supply, and everything powers down at the same time. Otherwise, you'll have to live with "not quite fully compliant" I2C interfaces. I recall that the license agreement with Philips says that the interface can only be called I2C if it is fully compliant. Just my $0.02 worth, Allan. On 1 Jul 1998 03:56:53 GMT, "Austin Franklin" <darkroom2@ix.netcom.com> wrote: >tom curran <tom_curran@memecdesign_dot_com> wrote in article ><3597dfd3.593512078@mdsXboston>... >> Simple vs. Complicated for an I2C interface is all a matter of >> implementation. If you want a fully compliant and robust interface, >> complete with bus arbitration, you will use much more than 20 CLB's. > >We did it in 24 CLBs in a 3142A and it is fully compliant (what ever that >means...but it works with Dallas, National and Philips peripherals) and has >arbitration for multiple masters (and that works too). So, it does depend >on how you implement it ;-) > >Austin Franklin >darkroom@ix.netcom.com >Article: 10921
nice summary - thanks. one of the few usenet messages i've ever bothered to print. evanArticle: 10922
I've managed to get a bit further with this. I ran up the offending simulation as a VHDL testbench, under ModelSim. The results now are: 1 Real device - doesn't work, has a timing problem 2 Foundation simulation (and timing analyser) - reports no problems 3 ModelSim - reports a timing error. This simulation is pretty simple - I write six 16-bit words into the device and, some time later, I get a string of 12 processed bytes out. This part of the device is also pretty simple - an input register loading a DPRAM, counters for the ram read and write addresses, some muxes on the ram output feeding two 4-bit registers, which are then combined into an 8-bit output register. There's no fancy clocking (everything, including the DPRAM, from the main 15.6ns clock), but all the stages have clock enables. I can't duplicate the simulation exactly on the bench, but the ModelSim failure does look very similar to the real device failure - ocassional bad data at the output. According to Modelsim, one of the DPRAM outputs ocassionally has a nasty glitch on it. The Foundation simulation *doesn't* show this. At the same time in the 2 simulations, Aldec reports a stable high level on the appropriate bit, but Modelsim reports a long (6ns) glitch (1 -> X -> 1). Unfortunately, this isn't the end of the problem. The functional VHDL simulation of the DPRAM produced by logiblox couldn't produce this glitch (the inputs couldn't result in the outputs). The simprim (timing) model is producing the glitch, but I don't have enough of the VITAL sources to work out why, and there's nothing obvious in the SDF data which could result in a glitch with this timing. I'll post the timing and the DPRAM ins/outs if anyone's interested. So, my current feeling is that the glitch found by Modelsim isn't 'real', and isn't responsible for the real device's failure, despite the similarity. It looks like a problem with the simprim cell, and I've got this one with Xilinx at the moment. Evan (ems@nospam.riverside-machines.com)Article: 10923
Hi, I have a minor problem. Let's suppose I have a simple PCB, it contains a SRAM based FPGA (Altera Flex 8K), an EPROM, a single LED and some resistors and capacitors. I measure the current consumption of the whole PCB. The FPGA smoothly boots up and does nothing, the LED shows that the FPGA finished the boot-up procedure. Ideally the card takes up about 1.5 mA. After RESET I can see the following on my multimeter: - the consumption is about 4 mA, and slightly growing. - after 15 seconds (!) or so the consumption reaches 7-8 mA, then drops to 1.5 mA and stays constant. The same can be observed with a PCB that contains a Hitachi H8/300h uC, an EPROM, a quartz and the LED. Current consumption is of course higher, but it changes the same way as described above. I simply don't know what causes this phenomena (every unused pin is configured for output). I've already asked mailto:sos@altera.com but they didn't answer (although they used to). So if someone has ever met something similar, please let me know, what's the reason for it. Thanx in advance, and cheers Botond -- Botond Kardos - at Innomed Medical Inc. in Hungary eMail: Kardos.Botond@hu.innomed.NOSPAM phone/fax: (0036 1) 351-2934 fax: (0036 1) 321-1075 To get my real address just put the domain name in reverse order and remove 'nospam'. x@1.2.3 -> x@2.1Article: 10924
Botond Kardos wrote: > After RESET I can see the following on my multimeter: > - the consumption is about 4 mA, and slightly growing. > - after 15 seconds (!) or so the consumption reaches 7-8 mA, then > drops to 1.5 mA and stays constant. > It has to be a floating node somewhere, inside the chip or outside.The only phenomena with that kind of time constant are: 1.thermal and 2. floating nodes. I think we can eliminate thermal problems, since heat causes the current to go down ( in CMOS ). Floating nodes can lead to internal contention and can thus draw a lot of current. Now the question is: Where is the floating node? Just my $0.02 worth. Peter Alfke, Xilinx Applications
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