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
FPGA `98 Call for Papers 1998 Sixth ACM International Symposium on Field-Programmable Gate Arrays DoubleTree Hotel, Monterey, California February 22-25, 1998 http://www.ece.nwu.edu/~hauck/fpga98 ========================================================== As Field-Programmable Gate Arrays become more essential to the design of digital systems there is increased pressure to improve their performance, density and automatic design. This symposium follows the largest ever gathering of this kind last year in Monterey at FPGA `97. For FPGA `98, we are once again soliciting submissions describing novel research and development in one or more of the following (or similar related) areas of interest: FPGA architecture: logic block & routing architectures, I/O structures and circuits, new commercial architectures. CAD for FPGAs: placement, routing, logic optimization, technology mapping, system level partitioning, testing and verification. Interactions: between CAD, architecture, applications, and programming technology. Fast prototyping: for System level design, Multi-Chip Modules. Applications: use of FPGAs in novel circuits, as emulators and compiled accelerators. Field-programmable interconnect chips and devices (FPIC/FPID.) FPGA-based compute engines. Field-programmable analog arrays. ========================================================== Authors should submit 20 copies of their paper (12 pages maximum) by September 26, 1997. Notification of acceptance will be sent by December 1, 1997. The authors of the accepted papers will be required to submit the final camera ready copy by December 15, 1997. A proceedings of the accepted papers will be published by ACM, and included in the ACM/SIGDA CD-ROM publications. All submissions should be sent to: Sinan Kaptanoglu FPGA `98 Actel Corporation 955 East Arques Avenue, Sunnyvale, CA 94086 USA e-mail:sinan@actel.com phone: (408) 522-4319 fax: (408) 522-8041 ========================================================== General Chair: Jason Cong, UCLA, Financial Chair: Carl Ebeling, U. of Washington, Program Chair: Sinan Kaptanoglu, Actel, Publicity Chair: Scott Hauck, Northwestern U. ============================ Technical Program Committee: ============================ Michael Butts, Quickturn Jason Cong, UCLA Eugene Ding, Lucent Carl Ebeling, U. of Washington Scott Hauck, Northwestern U. Dwight Hill, Synopsys Brad Hutchings, BYU Sinan Kaptanoglu, Actel David Lewis, U. of Toronto Fabrizio Lombardi, Texas A&M Jonathan Rose, U. of Toronto Rob Rutenbar, CMU Malgorzata Marek-Sadowska, UCSB Gabriele Saucier, IMAG Martine Schlag, UCSC Tim Southgate, Altera Steve Trimberger, Xilinx John Wawrzynek, UCB Martin Wong, UT at Austin ============================================================================ Sponsored by ACM SIGDA, with support from Actel, Xilinx, Altera, and Lucent. ============================================================================ +-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+ | Scott A. Hauck, Assistant Professor | | Dept. of ECE Voice: (847) 467-1849 | | Northwestern University FAX: (847) 467-4144 | | 2145 Sheridan Road Email: hauck@ece.nwu.edu | | Evanston, IL 60208 WWW: http://www.ece.nwu.edu/~hauck | +-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+Article: 6576
PRESS RELEASE 06/02/97 10:00 Fintronic releases Super FinSim. Aggressive pric- ing! San Mateo CA (June 2, 1997) Fintronic USA, Inc. the supplier of high quality Verilog Simulators (with over 800 Verilog simulators installed worldwide) announces today that it has introduced Super FinSim to its product line. Super FinSim supports the entire Verilog language, as well as a complete simulation environment based on PLI, SDF, and VCD. Super FinSim provides any mixture of compiled and interpreted simulation in order to achive the fastest turnaround time. Short simulations are run in interpreted mode whereas long simulations are run in compiled mode. State- ments that are seldom executed are interpreted whereas statements often executed are compiled. Technology: With Super FinSim Fintronic has introduced the first mixed event and cycle simulator for Verilog. A smart partitioner analyzes the design and decides which areas can be simulated by the Enhanced Cycle Simulation (ECS) kernel. The rest is simulated by the event driven kernel. Therefore Super FinSim does not re- quire any special design methodology. It preserves accurate tim- ing information, can handle modules with path delays, and sup- ports simulation values of Xs and Zs. Super FinSim supports all popular display tools including Design Acceleration's SignalScan, Veritools's Undertow, and IK Technolo- gy's Ishizue. FinSim, which does not include the ECS kernel nor the optimizations made possible by the detailed analysis required for the ECS partitioning, is the simulator used in the Veribest Design System. As an option, Super FinSim supports FinCov a tightly integrated code coverage with very low overhead (under 35%), making it a vi- able solution for regression code coverage. Super FinSim supports co-verification of C code and hardware by providing an efficient interface between the simulated C code and the hardware simulation. Pricing and Availability: Super FinSim provides the best value for the money. List prices include one year maintenance. Prices on the PC platform have recently dropped. The new line of Super FinSim simulators range in price from $799 to $28,000 depending on product configuration and platform. Fintronic has a Web page at http://www.fintronic.com, which can be used for placing or- ders, requesting demo licenses, checking prices, etc. Fintronic provides hotline support and software distribution via the Inter- net. For more information send e-mail to info@fintronic.com or contact Dr Alec Stanculescu by calling 415 349 0108/x105. Mission: Fintronic USA has a mission to supply the highest per- formance Verilog simulators available at the best price/perfor- mance ratio. Fintronic will exhibit at DAC 97 in booth no 546 in the Anaheim Convention Center, June 9-12, 1997. Fintronic's vendor presenta- tion will be held in rooms A6/A7 on June 9, at 9:20am.Article: 6577
The annual IEEE workshop on MEMORY Technology, Design, and Testing will be held August 11-12 in San Jose, California An abstract of the keynote address is attached. Further details can be found at <ftp://ftp.cs.tamu.edu/pub/fmeyer/service/mtdt97/abstracts.html>. We are having sporadic problems with our file system. If you have any trouble retrieving stuff, please just <mailto:fmeyer@cs.tamu.edu>. KEYNOTE ADDRESS Matching Memory to the Power of Personal Computers Dr. Richard Foss MOSAID Technologies, Incorporated ABSTRACT For almost a decade, the makers of DRAM ignored the growing problems of accessing data at rates matching the exponentially growing demands of new microprocessors. The increase in memory densities meant that the traditional form of DRAM, with features and performance unchanged since the mid-70's, actually worsened with each generation from a systems perspective. A present-day DRAM offers an eighth or less of its capacity as an active memory: the remainder of the chip is dead capacitance even in an active cycle. In addition, the data is not in a useful format and the analog mess of timing parameters is a system designer's nightmare. Patches designed to increase raw bandwidth helped only marginally at the cost of making the timing specifications almost unworkable. A major advance came with the thrashing out of a new approach to the format of DRAM in the JEDEC committee of the EIA. The Synchronous DRAM (SDRAM) represented the first significant re-think of the functions of a DRAM since the 4K. All address, control, and data inputs were referenced to a master clock, latency was introduced and made programmable and programmable burst sequences allowed the memory to handle specific processors and clock speeds. As a key feature, multi-banking increased the probability of data hits on open pages of data. Despite some early skepticism, the SDRAM and its Graphics cousin, the SGRAM, is now seen as becoming the dominant form of main memory chip by millennium. Inevitable, what both system and chip designers have learned from the SDRAM revolution is what should really have been done! The still mounting pressure for more effective (as opposed to raw and hence generally unusable) bandwidth also drives work on potential successors to SDRAM or to future variants and enhancements of SDRAM. Simply upping the clock frequency is not enough. It is necessary to optimize the memory system by having chips which meet the following needs: 1) High effective bandwidth by matching the control and address bus rates to data rates both within a chip (bank to bank) and from chip to chip using an efficient protocol. 2) Resolution of real-world bus timing issues at high clock rates without imposing high system level cost or power burdens. These goals need to be met while simultaneously allowing an expandable and versatile system, conforming to an open standard capable of being implemented in a variety of ways. An attractive resolution of these somewhat conflicting demands has been achieved by an industry-wide consortium under IEEE auspices as SL DRAM. Adapting what has been learned from SDRAM, it has added capability in the areas noted while simultaneously reducing system power demands by control-driven power management. The pattern of future standards for the Memory business is thus being set by world wide groupings of the major players in this Multi-Billion dollar industry. There are numerous issues still to be fully resolved, not least those of economical production testing. The address will highlight some of these key topics.Article: 6578
In article <33934047.3888020@nntp.netcruiser>, Stuart Clubb <s_clubb@netcomuk.co.uk> wrote: >Altera FLEX10K is not PCI compliant. Doesn't matter how much marketing >spin you put on it, or how many glossy adverts you throw at the >market, or how much html you use on your web site. >1 >The Clock pin capacitance breaks the spec. This is due to be changed, >so they will be compliant on this checkbox at some point in the >future. >2 >The pci_a core as it stands has some signals from the PCI bus driving >two pins on the FLEX10K device. That breaks the spec, and puts a >maximum 20 pF load on the pin. >Today, they are not compliant with the PCI specification, end of story Hey, don't have a cow man! It's nothing that a few series resistors can't fix. They have to be large enough so that the input impedance with the resistor would equal the reactance of a PCI single pin, but small enough so that outputs can still drive the bus with enough current (assuming any output pins are doubled up). Around 10-27 ohms should work fine- make sure there's no ground plane under the resistor (to reduce capacitance). -- /* jhallen@world.std.com (192.74.137.5) */ /* Joseph H. Allen */ int a[1817];main(z,p,q,r){for(p=80;q+p-80;p-=2*a[p])for(z=9;z--;)q=3&(r=time(0) +r*57)/7,q=q?q-1?q-2?1-p%79?-1:0:p%79-77?1:0:p<1659?79:0:p>158?-79:0,q?!a[p+q*2 ]?a[p+=a[p+=q]=q]=q:0:0;for(;q++-1817;)printf(q%79?"%c":"%c\n"," #"[!a[q-1]]);}Article: 6579
Hi, all, We are about to make a decision on purchasing a FGPA synthesis tool. The candidates are FPGA Express from Synopsys and Synplify from Synplicity. The target FPGA is Lucent OR2C40A (40k gates), and the design is about 100k in size. We would like to know which one can handle the size of the design well, which one will produce a better result, and which one will run faster. Your comments are really appreciated. Please send your comments to Jinan.Lou@tanner.com. Thank you in advance JinanArticle: 6580
Hi All I have released new version of alternate Verilog FAQ. I added following things. 1] A new "Technical Topics" section. This time information about state machine design is added. 2] Book section is modified. J. Bhasker's new book is added in the list. URL is : http://www.comit.com/~rajesh/verilog/faq/alt_FAQ.html Please email me suggestions. Rajesh rajesh@comit.com -- Posted using Reference.COM http://www.reference.com Browse, Search and Post Usenet and Mailing list Archive and Catalog. InReference, Inc. accepts no responsibility for the content of this posting.Article: 6581
Jinan Lou wrote: > > Hi, all, > > We are about to make a decision on purchasing a FGPA synthesis tool. > The candidates are FPGA Express from Synopsys and Synplify from > Synplicity. The target FPGA is Lucent OR2C40A (40k gates), and the design > is about 100k in size. > > We would like to know which one can handle the size of the design well, > which one will produce a better result, and which one will run faster. > Your comments are really appreciated. > > Please send your comments to Jinan.Lou@tanner.com. > > Thank you in advance > > Jinan Hi Jinan! Didn't you forget one tool: Exemplar Logic's Galileo? Yes, I work for Exemplar! No, I am not a marketing nor a sales guy, and I know what it means to synthesize 100K gate designs... I really want to encourage you to look at our new GALILEO EXTREME. We put into it a lot of new technology to speed up the synthesis and improve the results. Go and find some information on www.exemplar.com or just visit our booth at DAC and have a look at it. Endric -- ----------------------------------------------------------------------- Dr. Nils Endric Schubert endric@exemplar.com Exemplar Logic 815 Atlantic Avenue, Suite 105 Alameda, CA 94501 Tel.: (510) 337 3761 Fax.: (510) 337 3799 -----------------------------------------------------------------------Article: 6582
Stuart, > >The design I 'use' does not have this problem for target or master... ;-) > > Sorry Austin, I should have responded faster. I bow to your superior > engineering skills (to Xilinx). Have they offered you money yet? Well, funny you should ask that... Before they 'contracted' High Gate Design to do their PCI interface design, they were in negotiation with me to buy my PCI design (which, by the way, was done about a year before the High Gate design was...), but then, suddenly, they decided to go with High Gate... Well, I did beta site the High Gate version, found many many many bugs and misunderstandings of the PCI spec etc. I do not recommend using their design, except for a possible template with which you can re-use 'some' of, but be prepared to do a lot of work on it... They (Xilinx) called me a while ago and asked if I would tell them how I did my design....and I told them that I was interested in telling them, but since they poopooed me on a few things over the years, I wasn't really happy with them....so we'll see if they pony up! > I therefore summarise my comments to: > > Lucent is fully PCI compliant. > Xilinx is fully PCI compliant, if you redesign the core they sell you. Or use Austin's design ;-) > Altera is not PCI compliant. > > Deathly silence from 2610 Orchard Parkway. Hmm, maybe I should put a > press release out... Who'se at that address? Later, AustinArticle: 6583
Poor old Dick Foss. He appears unwilling to admit that SDRAMS (on which his company holds patents and from which he hopes to profit) have lost the war. Indeed the entire 53-line abstract carefully avoids mentioning the winner: Rambus. Not only does Foss make zero buxx from Rambus' complete and total victory, he loses face because he backed the wrong horse: first CTT, then SSTL, now DDR. So, with head firmly implanted in sand, he goes around giving misguided talks about the superiority of SDRAM while studiously neglecting to say that it's already lost. Perhaps he'll even sic his lapdog (PdM) on me for pointing out the bankruptcy of his abstract, and, one presumes, the talk itself. Watch for posts from mosaid.com and/or Kanata, Ontario. Mark Johnson Silicon Valley, CaliforniaArticle: 6584
John McGibbon wrote: > It could also be used for > reconfiguring serial flash proms from Atmel. > > John > john_mcgibbon@memecdesign.com If yo would like more information on Atmel ISP FPGA configuration memories check out http://www.atmel.com/atmel/products/prod182.htm and the AT17C FAQ at url http://www.atmel.com/atmel/products/fpga/fpga1.html. These parts now support both 5.0V and 3.3V (LV) ISP read and/or write. The protocols used to program the devices in-system can be designed in FPGA hardware *relatively* easily you will find the link below useful if you are interested in going that path. You can contact Atmel at configurator@atmel.com if you have any futher questions. Martin Mason martin@atmel.com AT17 Series Apps and Marketing Atmel Corp. and Uffe Tyrsted Toft wrote: >Take a look on the following homepage. You will find an I2C bus controller >a.o. > >http://www.acte.no/freecore > >Regards >Uffe Tyrsted Toft >------------------------------------------- >ACTE NC Denmark A/SArticle: 6585
How about comp.arch.fccm Those of us remotely involved in the industry know this term - right ??? Martin Mason martin@atmel.com FPGA/AT17C Atmel Corp.Article: 6586
Eric Fleischman wrote: > > If you are attending the Design Automation Conference, don't miss > DynaChip's demonstration of their Fast Field Programmable Gate Array. > These unique FPGAs use Active Repeaters to propagate signals with > extremely short routing delays. > > At the demo you will see: > > - Over 100 MHz system-level performance using HDL synthesis flows > - Short, predictable routing delays that are not affected by fanout > - Ultra high speed I/O structures that support up to 200 MHz clock > and data > - ECL, PECL and GTL interface levels > ... cut... At the UK's Annual Advanced PLD & FPGA Day on May 14, Xilinx reported implementing three designs that clocked at 200MHz "worst-case guaranteed over temperature and voltage". The essential technique was to keep all the high speed circuitry close together and as small as possible - lower-rate signals were propagated elsewhere on the IC. The presentation (but not the published paper) stated that the main difficulty was in providing EXTERNALS (eg PCB layout) that didn't compromise the very high rate signals which the IC itself was happy with. What Xilinx certainly did not offer was 200MHz across a whole FPGA. Is this what can be expected from DynaChip? Tim Forcer tmf@ecs.soton.ac.uk Department of Electronics & Computer Science The University of Southampton, UK The University is not responsible for my opinionsArticle: 6587
Hi all, Is anyone out there using the RIPP 10 Board? If so, do you have or know where you can get a template ACF file for the FLEX81188 chip in socket U5? Cheers, Barry Rising barry@dcs.rhbnc.ac.uk Research Postgraduate Dept of Comp Sci Royal Holloway University of LondonArticle: 6588
Hi all, (Who messed with my netscape preferences!?!) Is anyone out there using the RIPP 10 Board? If so, do you have or know where you can get a template ACF file for the FLEX81188 chip in socket U5? Cheers, Barry Rising barry@dcs.rhbnc.ac.uk Research Postgraduate Dept of Comp Sci Royal Holloway University of LondonArticle: 6589
Please distribute widely but delete after June 15, 1997 =================================================================== SUMMER STUDENT JOB Ricoh Silicon Valley has a summer research position for an exceptionally talented graduate student in computer science. The central task is to analyze implementations of image processing and pattern recognition algorithms on novel, parallel configurable supercomputer hardware. The ideal candidate will have experience in the following: General knowledge * Computer architecture, especially DSP architectures * Parallel computing * Theoretical Computer Science . analysis of algorithms . computational complexity * Configurable computing * FPGAs * Image processing and pattern recognition Specific skills * C/C++/UNIX * SUIF compiler toolkit * MPI (Message Passing Interface) TERMS The employment period is flexible and can accommodate the successful applicant's schedule, but is roughly three months and can start immediately. Salary is competitive. ELIGIBILIGY Applicants must have an appropriate visa, work permit or green card. BACKGROUND LITERATURE * "Increased FPGA capacity enables flexible, scalable CCMs: An example from image processing" by Jack Greenbaum and Michael Baxter, IEEE Symposium on Field Programmable Custom Computing Machines (FCCM 97) Napa Valley, CA, pp. 252--258 (1997). * "Configurable Computing" by John Villasenor and William H. Mangione-Smith, Scientific American, pp. 67--71 (June 1997). RICOH SILICON VALLEY The California Research Center of Ricoh Silicon Valley is a small research lab focussing on information and image processing, recognition and communication. It is close to Stanford University and numerous Silicon Valley landmarks, including University of California's Berkeley and Santa Cruz campuses. The culturally rich cities of San Francisco and San Jose are nearby, as are holiday sites such as the Pacific Ocean, Sierra Nevada mountains and Yosemite National Park. The lab's atmosphere is collegial and informal, yet highly challenging and professional. QUESTIONS Please consult our web site http://www.rsv.ricoh.com for general background and http://www.crc.ricoh.com/~stork/MLPSummerJob.html for more specific information. Feel free, too, to contact Dr. Stork, at the below addresses. TO APPLY Please e-mail an application letter, resume in clear-text ASCII (including relevant courses taken, papers, and your telephone number) as well as the names, e-mail addresses and phone numbers of at least two people familiar with your work. E-mail your application to stork@rsv.ricoh.com before June 15. (Put "Summer student position" in the message header.) Ricoh Silicon Valley is an Equal Opportunity Employer, and strongly encourages qualified women and minority students to apply. --Dr. David G. Stork Chief Scientist Ricoh Silicon Valley 2882 Sand Hill Road #115 Menlo Park, CA 94028-7820 stork@rsv.ricoh.com 415-496-5720 http://www.crc.ricoh.com/~stork ===================================================================Article: 6590
[ Enclosed is a review of last year's (1996) DAC in Las Vegas. If you see me at DAC next week in Anaheim, tell me what you think's hot and what's not, so I can include it in this year's review! :^) - John ] ------------- !!! "It's not a BUG, jcooley@world.std.com /o o\ / it's a FEATURE!" (508) 429-4357 ( > ) \ - / The Fourth Annual ESNUG/DAC Awards: _] [_ "Teamsters, Vegas & DAC '96" - or - "One Engineer's Review of DAC '96 in Las Vegas, NV, June 3-7, 1996" by John Cooley Holliston Poor Farm, P.O. Box 6222, Holliston, MA 01746-6222 Legal Disclaimer: "As always, anything said here is only opinion." [ Check out pg. 16 in this week's (June 17th) EE Times or "cooley.gif" at "http://techweb.cmp.com/eet/eda/graphics/" for the photo of the DAC freebies and their awards! - John ] Once I saw the AFL/CIO (the Teamsters) also having a conference in the same hotel as DAC, I had this weird sense that Vegas more their town than ours. These live-for-the-moment truckers drank like fish, would gamble away a month's pay in an hour at a blackjack table, and could keep an army of those "private nude dancers" busy in their hotel rooms until dawn. In contrast, my fellow more analytical engineers tended to drink lightly, only gambled a little money at the best odds (which we carefully calculated), and were way too afraid of AIDS to ever do anything more than look at a "private dancer." While truckers impress each other with the inside scoop on Jimmy Hoffa, we engineers "wow" each other by discussing undocumented Synopsys commands. Truckers get belly laughs from telling offensive jokes about every race, creed, gender and color; engineers get their kicks safely reading "Dilbert." Clever engineers fret over which hot new start-up to join; clever truckers try to figure out how to skim off of the Teamsters Retirement Fund. (Even marketing people see us as "different". For conference freebies, truckers got fun everyday guy stuff like 6-packs of beer, boxes of cereal, and condoms. We engineers only got cheap T-shirts, coffee mugs and puzzles...) Although, I'm quite happy as an engineer, there's part of me now wondering: "Is it too late to sign up for tractor trailer training school?" Anyway, on with the 4th Annual ESNUG DAC Awards! GREENPEACE AWARD: If EDA companies were living creatures, they'd be in one healthy eco-system. DAC grew 10 percent (to over 16,000 attendees) and the exhibit floorspace itself increased 25 percent. The exhibitor count jumped from last year's 152 to this year's 164. 32 of last year's companies didn't come back with their own booths this year -- only to be replaced by 37 new companies. Although i-Logix, Vista, and Harmonix moved on, most of the missing 32 (like Attest, Exemplar, and Silerity) were gobbled up by larger companies as part of the EDA food chain. A few did simple name changes like Intergraph becoming VeriBest; a few even mutated, like the RaviCad's consulting becoming Virtual Chips, a PCI-IP company, -- but most of the new 37 are companies with new ideas. "I don't know why they keep showing that stuff. After three years they've still never had a single customer outside of IBM." - A user commenting on IBM EDA still showing BooleDozer and EinsTimer at this year's DAC. "Jesus! We beg and grovel for a bloody $2 billion and these slimeballs just saunter on in with $7 billion!" - An EDA vendor commenting that the jewelry convention also in Vegas during DAC displayed $7 billion of merchandise. (The EDA industry as a whole only nets about $2 billion per year.) BIG CHICKEN AWARD: When Nanette Collins (who does PR for VHDL International) asked me to attend DAC's first Workshop for Women in EDA (as an observer), a little angel in my right ear said: "John, now you can share that Bill Clinton part of you which strongly supports Equal Opportunity. It'll be a teary-eyed moment to celebrate our human diversity." Then a little devil in my left ear said: "Are you kidding!?! Nanette wants blood for all that press about sparsely attended VIUF's. She knows the Rush Limbaugh side of you dislikes Affirmative Action and is more than clever enough to make her fantasy EE Times headline 'COOLEY KILLED BY ANGRY MOB OF EDA FEMINISTS' come true!" (Self-preservation being a stronger instinct, I chickened out. I later heard 101 women and two men attended: Bob Bellinger of EE Times and an unnamed weasel/headhunter who kept asking everyone for their business cards.) WORST FIRST IMPRESSIONS PARTY: Simplex Solutions took customers to see the movie "Mission Impossible"; those who went were left stranded at a distant theater with no taxi's. Four AT&T engineers groused: "The final kicker was when 5 Simplex people jumped in their car laughing and just drove away." "I guess the moral is, don't entrust your life to a CAD vendor." - Eric McCaughrin of SGI, who took a defective Veritools water bottle on a 3 day camping trip in the Grand Canyon. MOST UNORIGINAL "NEW" PRODUCT IDEA AT DAC: Cycle-based simulators. At least a dozen companies (SpeedSim, Frontline, Fintronic, Synopsys, Vantage, Chronologic, Pendulum, Cadence, Cadence Alta, Synopsys, Mentor (rumored), and CAE Plus) have or are working on one. It's like the R&D staffs from these companies all call the same psychic hotline for advice. FIRST 3-D DEMO AT DAC: Mimicing the crowds of people in the 1950's who went to watch 3-D movies, nearly 300 DAC attendees donned special polarizing glasses to watch LogicVision's Built In Self Test (BIST) demo -- a DAC first. VERILOG VS. VHDL (PART 2): This time it's analog. The Verilog-A LRM was approved at this DAC. Cadence, Meta-Software and Apteq are alreay finalizing their Verilog-A simulators. (Nanette wasn't too happy informing me that the VHDL-A LRM, which was supposed to be done by this DAC, was again delayed.) "I still stand by those words." - Cadence CEO Joe Costello about his IVC quote that VHDL "was a $400 million mistake." (Ironically, Cadence also sells a VHDL simulator that has done quite well in the public benchmarks.) IP WAS HOT: Companies like VLSI Libraries, 3Soft, Technical Data Freeway, Virtual Chips, SAND, CAST and Synopsys were all getting attention from customers wanting re-usable designs. Not wanting to be left out, Xilinx pushed its LogicCore program, Altera released its LPM to the public domain, and Crosspoint touted its CoreBank program for third party designs. (There were even hot rumors that these companies were in the act of forming some sort of "IP Alliance" at DAC.) Some even saw Casacade as an IP company. FUNCTIONAL VERIFICATION WAS HOTTER: Eagle Design, DS Diagonal Systems, Chronology, Cadence Alta, NuThena, and even the tiny Levetate got special consideration from engineers trying to verify designs. But what most users ranted about was the newbie InSpec. While most companies in this category (which was easily confused with HW/SW Co-design) pretty much offered waveform viewers and a framework all wrapped up in a glitzy GUI, InSpec gave users an automatic way to generate functional test vectors combined with a clever graphical analysis tool that quickly points out coverage holes. System Science also offers a dialect of Verilog that makes test generation easier. MOST IMPROVED COMPANY: ViewLogic. They're no longer hemorrhaging employees (in fact, they're growing), the legal battle with the old Chronologic staff is *finally* over, and instead of giving out cheesy white T-shirts, this year they gave out golfer's putters which tied with Altera's basketballs for BEST DAC FREEBIE. View/Silerity's datapath compiler can now use accurate View/Quad MOTIVE timing plus interface to the third party Synopsys Design Compiler. View/Chronologic's VCS got ASIC sign-off status from Motorola, LSI, Toshiba, and Lucent. VCS Roadrunner is beefed up 5X to 10X plus it's now integrated with the View/Quad MOTIVE static timing analyzer and the View/Sunrise ATPG tools. View/Chronologic's universally-executable-yet- still-encrypted Verilog Model Compiler (VMC) is perfectly poised to take advantage of the growing IP boom. Good rebound, Viewlogic! "We started the user group and let it go. Like many fledglings, it struggled for a bit and died. I even think Sean Murphy was the Chairman the user's group at the time." - Viewlogic CEO Alain Hanover responding to Sean Murphy's asking about needing a user group to develop innovative tools. (Sean was never involved with Viewlogic's users group.) "To Alain Hanover, CEO of Viewlogic: What happened to the IC PowerTeam concept? I don't hear about it anymore." - Daniel Payne, Mentor Graphics, who was the Viewlogic marketing manager for IC PowerTeam before joining Mentor. "YES, YOU CAN TEACH AN OLD DOG NEW TRICKS" AWARD: Unlike Savantage's SavanSys, which is more of an engineer's business analysis tool to juggle costs and a system's physical implementation (by answering: "Would this work better as 5 small PCB's and a backplane or as 2 bigger PCB's with connectors?"), Omniview's FIDELITY is truely a system-level synthesis tool. An offshoot of the Carnegie-Mellon "Micon" project, it works as an add-on to Mentor's Design Architect and Viewpoint Editors. The user enters in a block diagram a very detailed, highly parameterized design which FIDELITY then breaks down into specific commercial IC's. It's a lot like schematic capture on steroids and with a slightly higher IQ. POWER! POWER! POWER!: The power freaks at this year's DAC got off on EPIC's AMPS (a swapper that could optimize 30,000 transistor designs based on power, delay and area), Simplex (for their circuit-level power analysis tools that shows IC hot spots), Sente's WattWatcher (an RTL-level pre-synthesis power analysis tool), System Science's toolset, and Cadabra's LILA. Synopsys Power Compiler got some interest, but there were far too many companies with SPICE level/mixed signal solutions to choose from at DAC: Agape, Anagram, Analogy, Ansoft, Bell Labs, CAD-Migos, Cadence, Contec, Deutsch Research, Interactive Image, Mentor, Meta-Soft, Microsim, OEA, Symetry, TMA, and VeriBest. "After 15 years in this business, you stop believing in physics." - Gary Smith of Dataquest, after a professor just explained how it's physically impossible to design at 0.18 microns. "We call them 'Cheech & Chong'." - An EDA competitor referring to Cooper & Chyan Technology. MOST DRAMATIC NEW PRODUCT ANNOUNCEMENT: The Friday before DAC, Simon Perry, Chief Editor of a British electronics publication phoned around in the EDA community asking "What's Synopsys' big secret annoucement at DAC?" On the Free DAC Monday, Synopsys had a security guard standing next to a workstation covered with black drapes and a big "?" sign. There was an extra sign saying "You can find out at 2:30 today." At 2:30 they had over 200 people standing around this draped workstation. Moments before the big Synopsys 2:30 press announcement, the Synopsys people even caught and kicked out Synplicity CEO Alisa Yaffa while she protested: "Hey, this is a press announcement!!! It's public information!!!" Turns out that Synopsys was annoucing FPGA Express, a PC based FPGA synthesis tool that will directly compete with Synplicity. (Oddly enough, I thought I heard Nanette whisper "YES!" when they said FPGA Express had a built-in VHDL tutorial but no Verilog tutorial.) "Seventy percent of my sales are on the PC. Seventy percent of my revenues are from UNIX." - ViewLogic CEO Alain Hanover two years ago justifying why he bought so many top of the line UNIX-based EDA companies. "Take a gamble on us! SYNOPSYS" - The logo on the freebie pocket protectors Synopsys gave out at their DAC party four hours after announcing FPGA Express. "I'd be happy to provide them the catsup for the crow they're now eating." - Dan Ganousis of VeriBest, who pushed EDA on Windows NT for 3 years, noticing how the big EDA companies are now jumping in. MOST CURIOUS NEW COMPANY: TriQuest offers a finite state machine thingy that does all sorts of bizarre optimizations, decompositions and transmogrifications on RTL level state machine code. Whoa! "NICHE WITHIN A NICHE" AWARD: Promoting a funky sort of BIST, CrossCheck sells the ability to make low power designs, fabbed only at Asian foundries, fully testable. Frequency Technology offers a sub-half-micron, interconnect calculator that specializes in troublesome 3-D geometeries. Two runners up: Incases makes a 3-D radiation simulation tool for PCB design and K2 offers "automated reticle synthesis and high speed viewing." MOST LIKELY TO GET SUED: While other EDA vendors are trying to get Synopsys to put their proprietary compiler directives and pragmas into the public domain, ACEO Technology is already openly using them. MOST APPROPRIATE FREEBIE: Cadence gave away free stopwatches so engineers could time, in minutes, how long they still had a job once a Cadence Spectrum Services salesman found their VP of Engineering. You can even set it to time in seconds once the Spectrum Technical Assessment Team arrives at your site! Cool! "OK, Spectrum Consulting may not be for everyone. No one is making you use it. I've never held a gun to anyone's head... yet." - Cadence CEO Joe Costello responding to John Cooley's pointed questions about his consulting division. BEST DAC PARTY: A tie between Quickturn and Mentor/Sun/HP. Users loved the Monday night party where Quickturn took everyone to see Penn & Teller's comedy/magic show. A lot of them enjoyed the quips Penn (the one who speaks) took at engineers and Quickturn. One such Penn quote: "Welcome to Quickturn. This is all mirrors, deception, misdirection and slight of hand... Oh, I'm speaking for Penn & Teller, not Quickturn right now." Teller (the one who never speaks) even spoke to answer questions! On Tuesday night the hot party to crash was Mentor Graphic's 15th Anniversary (which was also co-sponsored by Sun and HP.) They rented the entire MGM Theme Park where users got to see an acrobatic dualing pirate show, go on water rides and simulated underground rides, see all sorts of roving musicians and odd characters roaming the park (like clowns, giant mice, and Elvises) plus two free drinks and dessert. ("Beer and cake, yum!") Everyone got a stuffed dog doll and a nice Mentor/Sun knit shirt as freebies. "We had one influential Asian customer ask that we send both a male and a female stripper to his hotel room. Nobody knows what happened for those two hours -- but it cost my company $700." - An anonyous voicemail from someone claiming to be an EDA vendor responding to a survey question. "What's 'parasitic extraction' ?! Getting your ex-wife off of alimony?" - Overheard by a PR agent immediately after a press briefing. "WHEN HARRY MET SALLY" FAKING IT AWARD: Escalade. Most of the ESDA vendors competed in the "Great ESDA Shootout" at the last HP Design SuperCon. It had an impact on these companies. Wally Rhines, CEO of Mentor, said his System Architect is being changed to make it more useable because of what happened in the Shootout. After clearing up a reporting snafu that Speed Electronics did, I resynthesized Speed's design using FSM Compiler two days before DAC, getting results of 1.80 nsecs -- placing Speed up with the hand coders in the shootout. And Summit Design should get some sort of DAC FREQUENT FLYER BONUS AWARD for how much mileage they got out of doing so well in the Shootout. The kicker was that although Escalade chickened out of the ESDA Shootout, they apparently were telling customers on the DAC floor some incredible synthesis numbers they got from doing this shootout! MOST CONTENT-FREE (AND ANNOYING!) FLOOR SHOW: For the *second year* in a row, both users and EDA vendors felt the HP floor show was the most content-free in all of DAC. "I got so bored I didn't even wait around to get whatever freebie they were giving out." said on EDA user. The EDA vendors (especially any within a four booth radius) were especially pissed with HP because they had a very loud simulated earthquake every 20 minutes that was very annoying. CAN'T KEEP IT UP AWARD: Many were frustrated that DACnet was down for most of DAC. "This year it vasn't vorth hacking. It vas down so much nobody could use it.", said the Swiss president of RubiCAD, Michael Reinhardt, who hacked DACnet last year to send junk e-mail to potential customers. "Is there a technical person in the house?" - John Cooley to a room filled with hundreds of engineers when the microphones went dead on the DAC/CEO panel. "It's also an early warning system. If you're at work and you're seriously thinking of opening and eating it, it's time to go home." - Michael McClure of Chronology, which gave away humorously relabelled cans of SPAM winning the FUNNIEST FREEBIE AWARD. MOST "OUT OF SYNC" DAC PARTY: The official DAC "Black & Blue" biker party was kept quite true to theme. They gave everyone black T-shirts with a red flaming DAC logo, they had lots of good food, an open bar that kept serving until about midnight, and a band that played very dance-able music. (I even saw Nanette dancing with her husband while she was wearing her "VHDL Uber Alles!" T-shirt.) They even had a temporary tatoo parlor plus vintage Harlies occassionally going through the room. The only problem was this theme had virtually *nothing* to do with Las Vegas -- but it would have fit in *perfectly* with last year's DAC in San Francisco where various subcultures there truely love leather, tatoos, and men on motorcycles! INTERESTING ODDS & ENDS: Design Acceleration, Veritools, Simutest, and InterHDL offered Verilog tools like waveform displays to source code "lint" programs to an optimizer for VCD files to queueing tools to Verilog/VHDL translators. On the VHDL side, LEDA offers another set of tools including a VHDL encypter and Intellx demo-ed a VHDL analyzer. Chronology's TimingDesigner did automated timing diagrams and analysis on RTL-level code. Design Acceleration, Veda, and Simulation Technologies are all plugging Verilog/VHDL code coverage tools. The SysAdmin types liked Platform Computing's LSF, a workstation load balancer which helped users schedule EDA tool runs over networks, Runtime's VOV, a tool that graphs dependencies between thousands of files in a project, and Spectra Logic's carousel that juggles either 60 4mm DAT's or 40 8mm cassettes to automatically back-up 240 to 280 gigabytes of disk. "I dunno. I thought it said 'VeriCow' the first time I read it." - A user reacting to Simulation Technologies sign announcing their "VeriCov" coverage tool. "Our simulators are so hot, they knocked him out!" - Dan Ganousis, VeriBest, whose DAC booth manager passed out from heat exhaustion 10 minutes before the DAC floor show started. TESTING FOR THE TRUELY PICKY: Not to be confused with wishy-washy functional testing, these companies focus on finding *all* the bugs in a design. ATG plugged INTELLECT, a killer partial scan technology that can handle gated and derived clocks, asynch logic, FF's, latches, RAM's and ROM's. Attest and Simucad are in the fault simulation software business. Synopsys, Mentor/CheckLogic, Viewlogic/Sunrise and Syntest were all doing ATPG demos. Intellitech, Syntest and LogicVision barked about BIST and BSDL. And still in the new paradigm department, Chrysalis displayed the industry's first interactive formal verification tool: Design Explore. "Formal verification isn't a nice thing to have on a project; it's an absolute must have!" - Alex Silbey, Silicon Graphics on the DAC verification panel. "OBJECTS IN MIRROR ARE CLOSER THAN THEY APPEAR" AWARD: Although many users feel Quickturn has the current lead over IKOS, Zycad, Aptix, Mentor/Meta- Systems, and Synopsys/Arkos in the special-hardware-for-hardware-designers market (especially since some of these aren't even on the market yet!), some are now openly noticing how the growing gate counts in FPGA's from Xilinx, Altera and Lucent might unexpectedly be changing this landscape. MOST SUICIDAL DAC FREEBIE: Interactive Image gave away 300 completely free, unrestricted copies of their $600 Electronic Workbench ($180,000 in total). Hasn't anyone told these guys EDA vendors come to DAC to *sell* tools? "By the turn of the century, you will have 5 devices on your body with their own IP addresses. Where you put them is your own business..." - Netscape CEO Jim Clark in his DAC Keynote address. "Yea, a few of us out dropped out of the EDA business to get into the Internet business. Less hype and more money." - Overheard from an ex-CrossCheck employee. THE SHAPE OF THINGS TO COME: The most immediate response engineers give is "higher levels of abstraction" so they talk about Synopsys's Behavioral Compiler and High Level Design System's RTL-level floor planner. Others speculate: "Because silicon is cheap, ASIC design is going to be replaced by embedded controllers -- so we best start learning to use HW/SW co-design tools like Mentor/Microtec's XRAY toolset or Hyperceptions's DSP-specific RIDE tool.", while many of the hype driven EDA developers aren't thinking specific tools as much as: "Hey, the Internet is hot! -- let's make JAVA based EDA tools that we sell on Web pages!" BACK TO THE "REAL" WORLD: After 5 days in mercenary Las Vegas, I was quite happy to get back home. The lights and glitz are fun, but it's not reality. After sleeping two days to recover, I knew I was back into my normal routine when I had started writing an EE Times column about the attendance numbers for the recent co-located IVC and VIUF conferences. Apparently over 650 attended the pro-Verilog IVC while less than 225 attended the pro-VHDL VIUF. (Nanette's going to scream bloody murder!) - John Cooley part-time EDA industry gadfly full-time contract ASIC/FPGA designer P.S. If you thought this review was on-the-money or out-to-lunch, please tell me. I love getting frank, honest feedback from fellow engineers. P.P.S. In replying, *please* don't copy back this entire article; a 14,400 baud modem attached to a 386 on a sheep farm can handle only so much! :^) =========================================================================== Trapped trying to figure out a Synopsys bug? Want to hear how 4258 other users dealt with it ? Then join the E-Mail Synopsys Users Group (ESNUG)! !!! "It's not a BUG, jcooley@world.std.com /o o\ / it's a FEATURE!" (508) 429-4357 ( > ) \ - / - John Cooley, EDA & ASIC Design Consultant in Synopsys, _] [_ Verilog, VHDL and numerous Design Methodologies. Holliston Poor Farm, P.O. Box 6222, Holliston, MA 01746-6222 Legal Disclaimer: "As always, anything said here is only opinion."Article: 6591
I want to be able to modify a xilinx 4020E at will. As the modifications are localised (always in the same CLB, and i just want to change the function generators), the fastest way is to change the bitstream. It could change the ".lca", but it's too slow (it takes a lot of time to recreate the bitstream and then send it to the FPGA). (the main target is fast reconfiguration..., so i cant wait 30 seconds or more between 2 reconfigurations) Here came my problems. It's easy to locate the position of the tables coding the function generators inside the bitstream. BUT, i must calculate the "4 Error Check Bits" comming at the end of each "data frame". (The DataBook says very few things on that) THEN ? Could you please send me details on -how to make the 4 control bits -what (and where) is inside the bitstream for a XC4000E. Thanks.Article: 6592
For dynamic changing of the bitstream as you describe, there are actually two possibilities. 1) Turn CRC off (an option in the makebits of your baseline bitstream) in which case, you only need to patch your function generator bits. You can see that the CRC field (last 4 bits of each frame) is the same 4 bit constant, and you dont need to change it in your modified bitstream (the code is 0110). You do need to run makebits with CRC off, not just set all the fields to 0110, as this wont work. The reason is that there is 1 more bit that needs to be flipped, to tell the chip that CRC is off. After this baseline bitstream is created, you can change it as much as you want, and the chip will eat it up and load it into the chip, no matter what the content. Be careful. 2) Modify the bitstream as you have suggested, and then modify the CRC field at the end of the frame. The way the CRC is calculated, requires that ALL CRC fields starting at the one that is at the end of the first frame that you modify, and continuing to the last CRC field in the last frame (which is 11 bits long). If you want to do this method, may I recomend a review of patent 5321704, which should be available from the excelent IBM patent server on the WWW. Note: The CRC field provides a check of your storage media for the bitstream (prom, eeprom, disk, RAM, core, notches on a stick ...) and the mechanism by which you transmit the bitstream to the FPGA. In well designed systems that have reliable storage, and well designed configuration hardware (failures are usually caused by poor signal integrity design) , CRC detected errors do not occur. The CRC does not check that the bitstream is functionally correct, but since the only way to create a base line bitstream is with makebits, and it performs a DRC on your design while creating the bitstream with CRC, you are pretty much guaranteed that the bitstream is ok. When you start playing with the bits, you are on your own. If you only change function generator bits, you are pretty safe. Start playing with bits that control routing and you can create designs with internal contention. For your application I suspect that my first suggestion is what you should do. Philip Freidin. In article <33959368.167E@pam.devinci.fr> Cyril Muller <muller@polytechnique.fr> writes: >I want to be able to modify a xilinx 4020E at will. >As the modifications are localised (always in the same CLB, and >i just want to change the function generators), the fastest way is >to change the bitstream. > >It could change the ".lca", but it's too slow (it takes a lot >of time to recreate the bitstream and then send it to the FPGA). >(the main target is fast reconfiguration..., so i cant wait 30 >seconds or more between 2 reconfigurations) > >Here came my problems. > >It's easy to locate the position of the tables coding the function >generators inside the bitstream. >BUT, i must calculate the "4 Error Check Bits" comming at the end >of each "data frame". (The DataBook says very few things on that) > >THEN ? > >Could you please send me details on >-how to make the 4 control bits >-what (and where) is inside the bitstream >for a XC4000E. > >Thanks.Article: 6593
Viewlogic will be presenting the Advanced FPGA Design Demonstration at DAC. The ADVANCED FPGA DESIGN demonstration highlights: *** ViewSynthesis as a totally redone and highly optimized FPGA synthesis tool. ViewSynthesis supports 100% Synopsys compatible VHDL language and has vendor specific optimizations for excellent use of resources and speed optimizations. It's new GUI features one button use with preset vendor specific defaults, but still offers controls for advanced users. *** IntelliFlow and ViewSynthesis in taking a mixed schematic and VHDL design from start all the way through place and route, final post-route timing simulation, and. creation of a PCB symbol and timing model. The demonstrator will be able to select parts and run flows for Actel, Altera, Lattice, Lucent, Vantis, or Xilinx. *** The design being used contains a state machine and data path to highlight the various optimization capabilities of ViewSynthesis. The design is created using ViewDraw for a top-level block diagram which instantiates four separate VHDL modules. *** The demonstration shows that Viewlogic has integrated our FPGA tools along with the vendor place and route tools into an easy to use methodology (flow) without taking away any of the control from the designer. It targets all FPGA designer, both infrequent users as well as the most advanced, since it uses default settings which will generally give the best results first time through, but gives the most experienced designers the ability to access every command switch in every tool for ultimate control. The demonstration shows this control by setting different ViewSynthesis control settings for the four different VHDL modules in the design. Please stop by booth #1751 for a demonstration! Thanks. Adam J. Elbirt Senior Programmable Solutions Engineer Viewlogic Systems, Inc.Article: 6594
In article <MJOHNSON.97Jun3205539@netcom12.netcom.com>, mjohnson@netcom12.netcom.com (Mark Johnson) writes: |> Poor old Dick Foss. He appears unwilling to admit that |> SDRAMS (on which his company holds patents and from which |> he hopes to profit) have lost the war. Indeed the entire |> 53-line abstract carefully avoids mentioning the winner: Rambus. |> Not only does Foss make zero buxx from Rambus' complete and |> total victory, he loses face because he backed the wrong horse: |> first CTT, then SSTL, now DDR. So, with head firmly implanted in |> sand, he goes around giving misguided talks about the superiority |> of SDRAM while studiously neglecting to say that it's already lost. |> |> Perhaps he'll even sic his lapdog (PdM) on me for pointing |> out the bankruptcy of his abstract, and, one presumes, |> the talk itself. Watch for posts from mosaid.com and/or |> Kanata, Ontario. |> |> Mark Johnson |> Silicon Valley, California The aclaim for Rambus is by no means universal. I guess that makes IBM a lapdog for Dick Foss? Just out of curiosity, what is the standby current of say, 10 GB of RDRAM?. I don't see any of the server companies lining up behind Rambus, unless one counts Intel as a server company. And I haven't seen anything from them besides an article in EETimes. Do the new Pentium II chipsets use RDRAM? We're a little isolated here, so clue us in. What is the story out there in silicon valley? -- Del Cecchi Frozen Tundra, Minnesota Opinions mine all mine. ps Who are you working for? Or do you just have Rambus stock? Mighty vitrolic attack for a normally staid technical newsgroup.Article: 6595
Hi, I was looking for a free copy of a vhdl-to-verilog translator. Does anyone know of any such thing available on the net. Pl. reply to my address knicks@corp.cirrus.com ThanksArticle: 6596
Let me see, you signed your name and your company and your domain and you are in the MN phone book and you supplied your email address, yep you are correct you did supply that information. I just wanted to clarify your identity without users having to search for your name, company, domain, email or the MN phone book in order to be aware of that particular piece of un-obvious information. Now, if that is a personal attack then you are a pretty sensitive kind of guy. Regards, Ben Blair Manager Field Applications QuickLogic Corporation The above comments are my own and in no way express the views of QuickLogic Corporation. (Typical legal disclaimer) :^) In article <19970531210528300253@1cust93.max1.minneapolis2.mn.ms.uu.net>, john@customer1st.com (John Sievert) wrote: > > Nothing hidden here! > > Let me see, I signed my name, my company name is in my domain name, both > my name and my company's name are in the phone book, and the POP I > connected from is listed right in the email. What's wrong with being > the sales engineer here? Aren't you a QL employee? Aren't we all > involved in this industry? > > The issue isn't who works for who, but what's in the summary judgement - > and therefore, who owns what. That isn't mud slinging, its a fact and > now a matter of public record. How is that mud slinging? What's your > problem? > > If you've got something to say on this case, I'd be interested in > hearing it. This court case is complex, but interesting and worthy of > discussion. If you want to make personal attacks, I guess I'd consider > that a waste of time. But, if that's the image you want to project for > you and your company, I guess that's your business. > > Regards > John Sievert > Customer 1st, Inc. > > <Blair@QuickLogic.com> wrote: > > > In article <19970526224329123272@cust4.max1.minneapolis.mn.ms.uu.net>, > > john@customer1st.com (John Sievert) wrote: > > > > > > Technology, which it appears, probably came from Actel (per summary > > > judgement against QuickLogic on patent infringement.). > > > > > > <kevintsmith@compuserve.com> wrote: > > > > > > > In this case, it's not marketing hype, just superior > > > > technology. > > > > > > -- > > > Regards, > > > John Sievert > > > > Aren't you the Actel Manufacturers Rep in the MN area? > > > > It would behoove you to not sling mud and try to hide your identity. > > > > Regards, > > Ben Blair > > Manager Field Applications > > QuickLogic Corporation > > > > -------------------==== Posted via Deja News ====----------------------- > > http://www.dejanews.com/ Search, Read, Post to Usenet > > -- > Regards, > John Sievert -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to UsenetArticle: 6597
I would like to make a simple PCI interface card that would allow me to do some basic I/O with 'outside world'. I found quite some information about the ISA interfaces, but basically nothing usefull about PCI. I would be very thankfull if someone could tell me how I should approach this problem (point me to some resources on net, or literature). Ivan Hamer (ivan.hamer@toronto.edu).Article: 6598
So let me get this straight... You had a working, fully compliant design before Xilinx even got off the ground with their buggy, half speed interface. They were "in negotiation" with you to purchase your design, yet decided that their future customers would be much better served by an unproven, non functional, non compliant product that needs a lot of hard work before it is usable. Naah. Doesn't sound like Xilinx's customer orientated approach to me! ;-) >They (Xilinx) called me a while ago and asked if I would tell them how I >did my design....and I told them that I was interested in telling them, but >since they poopooed me on a few things over the years, I wasn't really >happy with them....so we'll see if they pony up! May you squeeze them 'till the pips squeek! >Or use Austin's design ;-) Surely that should read "Or buy Austin's design."? Or are you about to become the ultimate PCI Philanthropist? >Who'se at that address? Altera. StuartArticle: 6599
On Tue, 3 Jun 1997 22:33:13 GMT, jhallen@world.std.com (Joseph H Allen) wrote: >Hey, don't have a cow man! It's nothing that a few series resistors can't >fix. They have to be large enough so that the input impedance with the >resistor would equal the reactance of a PCI single pin, but small enough so >that outputs can still drive the bus with enough current (assuming any >output pins are doubled up). Around 10-27 ohms should work fine- make sure >there's no ground plane under the resistor (to reduce capacitance). And this has been characterised and proven compliant? I'm not "having a cow" as you so eloquently put it. It just concerns me that some engineers are reading, and swallowing, every piece of "marketing" seen on a web site or pdf file. Marketing joke: A mathematician, an engineer, and an FPGA marketing VP are all asked how much two plus two is. The mathemetician says "It is exactly four point zero." The engineer says "Four. Now can I have a pay rise?" The FPGA marketing VP says: "What do you want it to be?" Stuart
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