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
Realistically, I doubt you'll be able to get any current FPGA with any real logic implemented within it to run at 100MHz. It's true, that you can achieve 100MHz operation with internal register to register logic, but not many designs I know only implement this. Perhaps if the only 100MHz operation you are looking for is an internal counter or something. You will achieve your best timing by pipelining, but my experience has shown that a "typical" large design can usually achieve anywhere from 25MHz to 40MHz operation depending on the vendor and amount of time spent optimizing your design. Note that 100MHz operation is even a difficult goal with most ASIC designs, and they can be optimized for speed. CPLDs are more predictable and can be faster in some cases, but again, I doubt you'll hit 100MHz as a general design rule. Do the math. Just look at your on-chip and off-chip signal times, your register setup, etc. and you'll see precious nanoseconds eaten up, and this doesn't even include your board delays and skews. In article <1996Jul18.115255.6511@srs.gov>, Bryan Bushart <bryan.bushart@srs.gov> writes: |> Marc Palmarini wrote: |> > |> > Hi, |> > |> > I'm currently working on a project wich involves very-high speed clock rate. Over the 100 MHz. |> > We're considering CPLD and/or FPGA. From what I can see, and have been informed by several compagnies. |> > CPLD offer less complexity, but more speed. Is this right ??? Also, I'd like to know if anybody |> > has had to work with different brand of CPLD/FPGA. We've been looking at two brands, first there is |> > the CYPRESS line, and then there's the the QuickLogic line (any comment/suggestion). Are there any other |> > lines that works at a higher than 100 MHz clock. Typically our project needs to access RAM (Burst Technique) |> > with acces to a peripheral (LCD Interface). Wich will need to be able to calculate an Adress/Data from an |> > (x,y) pixel position. That is about all the job for the PCLD/FPGA. I need at the very least 64 I/O line. |> > |> > Thanks in advance for any comments and/or suggestions. |> > |> > -- |> > ============================================================ |> > Tel:(418)-871-8977 Marc Palmarini |> > Fax:(418)-871-9021 ATI International |> > |> > E-Mail:amphi_t@acica.com |> |> Timing for CPLDs is deterministic and predictable however on FPGAs the |> timing depends on the place and route. Which will be faster depends on |> the size of the design. FYI there is no difference architecturally |> between Cypress and Quicklogic FPGAs. Another possibility is the FPGA |> line from Actel. If you need a reprogrammable FPGA you might want to |> look at the ORCA line from Lucent Technologies (formerly AT&T). |> |> ------------------------------------------------------------------------------------------ |> All opinions expressed is this message are my own and should not be |> interpreted as representing the views of DOE or WSRC. |> |> Bryan Bushart |> Net: bryan.bushart@srs.gov |> Phone: (803)725-3472 -- Mark Christensen ascom-Nexion email: markc@nexen.com 289 Great Road ph: (508) 266-2315 Acton, MA 01720Article: 3726
> >I am using Xilinx XC4000 FPGAs for my design that has few increment and decrement >by one and data bus is 8-bit integers. I was wondering whether there is any design that >uses less resources of FPGA than 8-bit adder/subtractors to increment an 8-bit number >by one. If there is such a method would anybody tell me what VHDL code infers this >hardware? Consider this: I once did a XC3064 design in which over 95% of CLB flip-flops were utilised. Moreover, this routed using old 1991 APR software (in about 20 hours!). This design was almost entirely made up of counters. Before and since, I have been doing "typical" designs and have never got anywhere near this level of utilisation. 70% is typical before the device won't connect-up, although when I recently bought XACT6 that made a big improvement. So, if you can do something with "counters" you will get a lot more into a device than if you do it with random logic. I would think that loading your number into an 8-bit loadable counter, and then giving that one clock, would be a lot more efficient than building an adder. With care, you could do the whole operation in 2 clocks. Peter.Article: 3727
Fernando Pardo wrote: > > Hi, > > I just installed the Xilinx kit for mentor graphics and I discovered > that there > is no Autologic synthesis library for Xilinx, is this an error?? I there > a public domain Xilinx synthesis library for autologic?? yes, i recently discovered the same thing. they are available at mentor's ftp "supportnet" site. the last time i looked, they were located in the following directory: /pub/mentortech/tdd/libraries/fpga if you have any problems finding the library you need there, i would suggest calling your local mentor rep or the mentor directconnect line (if its available in europe). good luck. -- _______________________________________________________________________ Lance Gin "off the keyboard Delco Systems - GM Hughes Electronics over the bridge, OFC: 805.961.7567 FAX: 805.961.7739 through the gateway, C43LYG@dso.hac.com nothing but NET!" _______________________________________________________________________Article: 3728
cgamrat@gaap.saclay.cea.fr (GAMRAT Christian) writes: >Message: 3 >From: cgamrat@gaap.saclay.cea.fr (GAMRAT Christian) >Date: 11 Jul 1996 12:19:29 GMT >Subject: What about the XC6200 ? > >Is There anybody out there who knows anything new about the Xilinx >XC6200 partially reconfigurable fpga ? The data available from Xilinx >has not changed for months and it looks like if they are not pushing >the device ahead anymore. I hope I'm wrong because this FPGA is a must >in the field of reconfigurable computing. It will be really Dommage ! >Thanks for any info on that . >Cheers, > >Christian. > > > ((((( > ( ) > ( 0 0 ) > ( ! ) > | \_/ | > |------------------------------ooO---------Ooo-------------------------------| > | De/From Christian Gamrat (MIND-1024) |> > | Commisariat a l'Energie Atomique / Atomic Energy Commission | > | Groupe Architectures Paralleles / Parallel Computing Group | > | DTA/LETI/DEIN/SLA/GAP | >O| CE-SACLAY 91191 Gif-sur-Yvette Cedex - FRANCE |>O > | e-mail : cgamrat@gaap.saclay.cea.fr | > | phone : +33 1 69 08 71 94 | > | fax : +33 1 69 08 83 95 |> > |----------------------------------------------------------------------------| In response wolf@unconfigured.xvnews.domain (William J. Wolf) writes: >Message: 1 >From: wolf@unconfigured.xvnews.domain (William J. Wolf) >Date: 12 Jul 1996 12:13:28 GMT >Subject: Xilinx reconfigurable logic strategy > >[was Re: What about the XC6200?] > >cgamrat@gaap.saclay.cea.fr (GAMRAT Christian) writes: >>Is There anybody out there who knows anything new about the Xilinx XC6200 >>partially reconfigurable fpga ? >>The data available from Xilinx has not changed for months and it looks like if> they >>are not pushing the device ahead anymore. >>I hope I'm wrong because this FPGA is a must in the field of reconfigurable >>computing. It will be really Dommage ! > >William Roelandts (new Xilinx CEO) stated that developing recongifurable logic >with "the ability to change the logic configuration of the FPGA during the >operation of the device" is something that will be aggressively continued. >Anyone from Xilinx care to expand on the details of this strategy? > >- -- >- - Bill Wolf, Raleigh NC >- - My opinions, NOT my employer's Also in response Andrew Morley <andym@trend.demon.co.uk> writes: >Message: 2 >From: Andrew Morley <andym@trend.demon.co.uk> >Date: Sat, 13 Jul 96 15:00:09 GMT >Subject: Re: What about the XC6200 ? > >In article <4s2rgh$eba@news.cea.fr> cgamrat@gaap.saclay.cea.fr writes: > >> Is There anybody out there who knows anything new about the Xilinx XC6200 >> partially reconfigurable fpga ? >... >> I hope I'm wrong because this FPGA is a must in the field of reconfigurable >> computing. It will be really Dommage ! > >Whilst it is an exciting offering, I don't see that it is that much more >radical than the Atmel CacheLogic parts. Perhaps they don't feel that thay've >got anything to bring to the party? I stand corrected if I am wrong. > >- -- > ----------------------------------------------------------------------------- >| Andrew Morley, Design & Development, Trend Communications Ltd, High Wycombe.| >| email: andrew.morley@trendcomms.com Phone +44 1628-524977 Bucks, UK.| > ----------------------------------------------------------------------------- Gentlemen, Please allow me to clear up some misunderstanding. 1) I'm glad Mr. Gamrat feels that the the Xilinx XC6200 "is a must in the field of reconfigurable computing." We at Xilinx couldn't agree with him more. Please rest assured that the XC6200 is very much alive. Information about the XC6200 on our WEB page is merely a glimpse of the 'architectural' benefits it provides the user. A complete datasheet is available directly from Xilinx, upon request. Currently, we are beta testing the software and sampling the silicon under NDA. Completion of a thorough beta test program for the XC6200 is a must before we release the product to the general market. I welcome those who believe that this architecture is a significant milestone in enabling reconfigurable logic applications, to participate in evaluating the software and silicon. 2) Mr. Wolf the details of W. Roelandts' Reconfigurable Logic strategy will unfold over the next year. Please stay tuned! 3) Mr. Morley the similarity of the Xilinx XC6200 with the Atmel Cache Logic parts begins and ends with perhaps the granularity of the devices. Both can be considered 'fine grain architectures'. Beyond this the XC6200 is completely different than the Atmel devices. a) The XC6200 is the first and only family of devices with a built in processor interface that allows direct random access of configuration memory as well as all the registers on the device. A must for coprocessor applications! b) The architecture supports partial dynamic reconfiguration down to the individual cell level. Elements of a single cell can be configured in as little as 40ns. c) It is a completely "open architecture", ie. everything the user needs to know to program and access the device is specified in the datasheet. This open systems approach is a key for any kind of 'computing' to take off. We feel this is also important for the 3rd party design tool market to take off. d) Configuration times for this architecture are faster than any other architecture (for equivalent gate densities) by orders of magnitude. e) Family includes devices ranging in gate densities from 9K gates up to 100K gates. f) Its the only architecture which can accept completely random bit streams without causing damage to the chip (has anyone had their CPU run off and write in areas of memory that it wasn't supposed to? how does one guarantee a non-random bit stream in such cases?) This is but a short list of the differentiating features. We believe the XC6200 architecture with its host of new and innovative features will be "Changing the Rules of System Design". A year ago we announced the "architecture" but most people read it as a "device" announcment and so expected the product to be released shortly. When the products did not start to appear, it was felt that the XC6200 was "slipping" or "not being pushed ahead anymore". In the meantime, work on the XC6200 family continues on or ahead of schedule and early indications are that this family will be a clear winner. ********************************************************************* * _ _ * * / /\/ / * * \ \ \/ Raj Patel XC6200 Product Line Manager * * \ \ Xilinx email:raj.patel@xilinx.com * * / / 2100 Logic Drive (408)879-4928 Tel * * / / /\ San Jose, CA 95124-3450 (408)879-4780 Fax * * \_\/\_\ * * * ********************************************************************* ----- End Included Message -----Article: 3729
>Klaus_Leiss@ccl.lhag.de (Klaus-Guenter Leiss) writes... >I used information downloaded in march from www.atmel.com. After >your last message i downloaded the information again. New File, larger >than the old ( it now contains also information for in circuit >programming), and as you stated the information regarding the polarity >fuse is now inverse to the old. After trying it with the new >information all went correct and i could also programm the polarity >fuse and not only the data block. Somewhere in an older thread someone >stated that Atmel changed the default polarity of its reset/oe fuse, >maybe at this time they changed also the programming algorithm. > Atmel has not changed the default Reset polarity or the device programming algorithim. The current on-line documentation available at http://www.atmel.com or Fax-on-Demand 1-800-29-ATMEL (Docs #602 & #603) is now up-to-date. Samples of the device can also be obtained through Atmel's home page. Technical questions can be sent to configurator@atmel.com Martin Mason. ------------------------------------------------------------------------- | Martin Mason | Snr FPGA/17Cxxx Applications Engineer | | Atmel Corp. | (email - me) martin@atmel.com | | 2325, Orchard Parkway | (email - help) fpga@atmel.com | | San Jose | configurator@atmal.com | | CA 95131 USA | (Tele) + (408) 436 4178 | | | (Fax) + (408) 436 4300 | ------------------------------------------------------------------------- |Need Atmel Lit. stuff ? WWW.ATMEL.COM or FAX-on-DEMAND 1-800-29-ATMEL | -------------------------------------------------------------------------Article: 3730
SORRY TO BOTHER YOU GUYS. I AM LOOKING FOR A ASIC DESIGNER WITH CELL3 OR CELL ENSEMBLE EXPERIENCE. FOR A JOB IN FLORIDA. THIS JOB WILL PAY VERY, VERY WELL FOR SOMEONE WITH THIS EXPERIENCE. IF YOU HAVE THIS SKILL OR KNOW SOMEONE WHO DOES, PLEASE LET ME KNOW ASAP. I WILL PAY A REFERRAL FEE FOR THE RIGHT FIT. PLEASE CALL AT 800-940-9401 OR E-MAIL vvirga@notes.techni-source.com. Thank You.Article: 3731
Hi, I will use two XC3064 FPGAs in a design in slave serial mode. What is the better approach to configurate the FPGAs: daisychain or by using a separate SPROM for each FPGA? By using the devices in daisy chain is it possible to download the configuration bit stream via the xchecker? Otherwise if i use the separate SPROMs, what about synchronistion problems? Thank you for reading. Marc Email: wilwert@crpht.luArticle: 3732
I've had four users independently suggest that I do an anonymous survey of engineers to see what we, as a group, think about these issues. As usual, I'm only interested in your answers, not who you are (other than the fact that you use these EDA tools) -- so your replies will all be treated as ANONYMOUS. Please reply *on* or *before* this Thursday, July 25th (three days from now) to "jcooley@world.std.com" so I can get this quickly tabulated and written up. - John Cooley P.S. After answering each specific question, please feel free to add your own anonymous comments on what the question addresses. 0.) Please indicate how Place & Route (P&R) tools affect you. Choose: [ ] I use them directly in my work. [ ] I support other engineers using P&R tools. [ ] I don't use P&R tools, but use related EDA tools for IC design like synthesis, etc. [ ] I develope/sell EDA tools. [ ] I don't use or support or make any EDA tools. Comments: 1.) "As an engineer, I believe that the American legal system is generally capable of rendering justice in technically complex lawsuits." Please answer TRUE or FALSE:___________ Comments: 2.) "As an engineer, if I were the judge in the Cadence/Avant! lawsuit, I ( WOULD / WOULD NOT ) grant Cadence's Sept. 11 request for an injuction to prevent the further sales of Avant! products that are alleged to contain Cadence technology." Please choose WOULD or WOULD NOT:____________________ Comments: 3.) "My primary sources of info on the Cadence/Avant! lawsuit came from (please "X", at most, your top 3 three sources): [ ] EE Times [ ] Internet discussions [ ] opinions heard at conferences [ ] opinions from co-workers [ ] opinions from boss/management [ ] direct interactions with Cadence [ ] direct interactions with Avant! [ ] other (please specify) _____________________ Comments: 4.) "Because of this lawsuit, I am (MORE LIKELY / LESS LIKELY / UNCHANGED) to do business with Cadence." Please choose MORE LIKELY or LESS LIKELY or UNCHANGED:___________________ Comments: 5.) "Because of this lawsuit, I am (MORE LIKELY / LESS LIKELY / UNCHANGED) to do business with Avant!." Please choose MORE LIKELY or LESS LIKELY or UNCHANGED:___________________ Comments: 6.) "Because of this lawsuit, I am (MORE LIKELY / LESS LIKELY / UNCHANGED) to do business with other P&R vendors." Please choose MORE LIKELY or LESS LIKELY or UNCHANGED:___________________ [ Other P&R vendors: Agape, Cascade, Compass, Cooper & Chyan, Fujitsu, Gambit, IBM EDA, Mitsubishi, Mentor Graphics, Sagantec, Sieko, Snake, SVR, Tanner, Timberwolf, VeriBest. ] Comments (& What Other Vendors In Particular?): Thanks for taking the time to answer this survey. - John Cooley Part Time EDA Consumer Advocate Full Time ASIC, FPGA & EDA Design Consultant =========================================================================== Trapped trying to figure out a Synopsys bug? Want to hear how 4599 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: 3733
I am using Xilinx 4000 series and I want to design a dual port RAM. I have got an application note that shows how to design it with 4000E series that CLBs are edge triggered.Now my question is how should I change that design in order to use 4000 series instead.Basically If I be able to make RM 16x1S that is available in 4000E libraries by RM 16x1 that I have in 4000 libraries now, my problem is solved. Any comment is highly appreciated. --- Sayed-Hossain Hajimowlana Dept. Of Elec. Eng. Office:Essex Hall Rm-213 Phone-ext:2590 Univ. of Windsor Home:253-9741Article: 3734
Felix K.C. CHEN asked about timing report from synthesis tools. In the Exemplar Logic Galileo tools, speed grades are set in the Output Options window for those technologies that have them. For those that don't, the Synthesis Options window contains options for degrading delay based on voltage, temperature and process. And as another poster mentioned, these delay numbers are estimates. In the spirit of any information is better than none at all, delay estimates can actually be helpful in the design flow. Larry Lapides Antares Corporation larry_lapides@antaresco.com *** Exemplar Logic synthesis, Model Technology simulation ***Article: 3735
In article <31F3E2D0.400B@crpht.lu> Wilwert Marc <marc.wilwert@crpht.lu> writes: >Hi, > > I will use two XC3064 FPGAs in a design in slave serial mode. What is >the better approach to configurate the FPGAs: daisychain or by using >a separate SPROM for each FPGA? By using the devices in daisy chain is it >possible to download the configuration bit stream via the xchecker? >Otherwise if i use the separate SPROMs, what about synchronistion >problems? > Thank you for reading. > Marc >Email: wilwert@crpht.lu If you have two XC3064s to configure serially, and plan to use the SPROMS, then they wont both be in slave mode. The lead device will be in master mode. You can load two devices in daisy chain with XCHECKER (in which case, both will be in slave serial mode). Use makeprom to make a combined bitstream file, and write it out in .TEK, .MCS, or .EXO format. XCHECKER can work with any of these. If you put both parts in master mode, each with its own SPROM, then they will configure independently, and as they will use their own internally and non identical clocks, the devices will not go done at exactly the same time. If your system depends on both chips going done at the same time, then daisy chain is a better possibility. With both parts in serial slave mode, you could provide an external cclk to both parts which would then cause them to complete at the same time (because they are both the same type of chip), but you can't use a free running clock, because cclk cant start to have transitions untill about 6 uS after the init pin goes high. If both chips are the same design (you didn't say), then you only need one copy of the program. Feed the bit stream to both chips at the same time, with one chip in master mode and the other in slave mode. Both will complete at the same time (because you also tied the cclk pins together). This will also configure in half the time which is almost always irrelevant. What I usually do is build the board to take both the XCHECKER (or my much faster and less capable parallel download cable) and what ever the system is suposed to use in production (SPROM or host load), with a jumper that changes the mode appropriately, and also flips a mux over for the source of data and cclk. Debug with jumper in, production with jumper out. Enjoy, Philip Freidin.Article: 3736
Thanks to all who replied. This is going to help a lot. As promised, here's the summary of replies, followed by my own results (or, more accurately, non-results so far :-). *********************************************************************** From: "J.Mawer" <j.mawer@sheffield.ac.uk> AMD palasm is now freeware, used to be a couple of hundred I think. This will do Mach devices upto MACH 230 as well as "ordinary" PLDs, you can get a copy from their web site. Mach XL 2.1 is also avaliable on their web site which can be used with upto Mach 465. Further information http://www.amd.com/html/products/pld/pld.html ********* From: Andreas Kugel <kugel@uni-mannheim.de> cplds: mach-xl from amd for their mach chips, free. www.amd.com cplds: ispLSI starterkit from lattice for the smaller ispLSI chips, free (or low cost). www.lattice-semi.com fpgas: motorola mpa devices (=pilkington fpgas), free place and route, free interfaces. front end tool may be self developed. www.pmel.com ********* From: Paul Walker <paul@walker.demon.co.uk> Best is PLDSHELL from Altera. Snap it up while it is still free and before it is subsumed into their other tools. They charge a little for the JTAG cable and the devices are quite expensive. But the tool lets you work on devices of any size, and the devices are quite nice, particularly the flash EPX880. ********* From: Gareth Jones <gareth@pmel.com> I saw your message on comp.arch.fpga, we (Pilkington Micro-electronics) provide free place and route tools for all our programmable technologies. Our FPGA's are currently sold under license by Motorola as the MPA1000 series devices. You can download the tools from our website (www.pmel.com) this site also has links to Motorola's site which contains information on device availability/pricing. The free version of the tools are fully featured but will only support devices up to 8000 gates. They contain free libraries for a wide range front end tools. An upgraded version of the tools supporting up to 22000 gate devices is available from Motorola. ********* From: wolf@aur.alcatel.com (William J. Wolf) >From previous consulting/contract experience: CPLD software is generally easy to get. FPGA software is a little tougher. Few customers need support for CPLDs. You will probably find more consulting/contract opportunities with FPGAs. Take a survey of local/regional customers to see which FPGAs they use, prior to investing a lot of time. Altera seemed more willing than Xilinx to give away software. This may depend on how well established you are, or their need for local support from third parties. Consider becoming a certified Xilinx trainer. This provides several benefits; another source of (modest) income, access to software, and bragging rights with customers. It does take a considerable investment in your time up front, with no guarantee that you will be "certified" and used. Xilinx provides honest feedback as you go to give you both a chance to bail. This only makes sense if you are serious about FPGA design, good, possess skills helpful in training, and are willing to travel to teach classes. ********* From: Ilpo Hamunen <ilpo@dna.fi> John, take a look at Lattice's homepage at http://www.latticesemi.com They have CD which is including a very nice piece of PLD software: Data I/O Synario Starter kit It will only support a limited number of devices, as the other packages that you have found, but it's for free and it's good ! ISP Synario Starter System software supports Lattice's ispLSI 2032 and 1016 high-density devices and all Lattice GAL devices. The software can be easily upgraded to support all or part of Lattice's high-density families. 1016 is a 44-pin, 64 macrocell device and 2032 also 44 pins but 32 MCs. The CD has also a lot of valuable documentation. ********* From: hjp@ee.umr.edu (Hardy Pottinger) Try PLDShell if you can still get it. Intel provided it for their FX780 series and Dave Van den Bout of Xess has a nice book out on it (FPGA Workout). Try ftp://ftp.vnet.net/pub/users/xess/FPGA_Workout_II/HTML/fpgawk2.htm Altera has it now and I don't know if they're going to support it or not or if it's even available. It was free so you may be able to find a copy in an archive somewhere. It's basically PALASM on steroids. ********* From: devb@vnet.net XESS Corp. sells low-cost CPLD hardware and software. You can check our WEB site: www.xess.com. You might also get a free copy of PLDSHELL from ALTERA. It supports a lot of their smaller PLDs and the EPX780 and 8160 CPLDs. Just call (408) 894-7144 for a free copy. *********************************************************************** I looked at the Motorola site, and the 14MB size was too much for my serial access! so I left the requested form for the free CD-rom, and am waiting for that. I also will order the Lattice CD-rom for $14.95 (I think that's the price -- I'll have to look it up again). I downloaded AMD's Machxl, and tried to ignore the silly procedure they give in the readme (copying from hard disk to floppy so I can install to the hard disk?!?!?), and the install procedure instantly jumped to the end screen when I started the install step. So I painfully copied all that stuff onto four floppies, and the exact same thing happened. By that time I was so exasperated I quit and didn't try PALASM, which has the same instructions. Anyone know what's wrong? I'm using a standard 486-40 with a Syquest EZ-Drive. I couldn't find PLDshell or any reference to it at Altera's web site, so I suppose devb's (is this Dave Van den Bout?) phone number is the way to get it. I'll check out Xess in the next few days. If anyone wants to hear my results, let me know. You can do it yourself, though.Article: 3737
We have two altera-designs we want to fit into xilinx devices. The first design is in Altera-HDL (.tdf-files) and the other design is written with PD-Shell (one .pds-file). We are working with Mentor Graphics Tools, Maxplus2 Version 6 and XACTstep 5.2 on Sun Sparc Station (SunOS 4.1.3). I tried to write an EDIF-netlist with the altera-tool (.edo) and to read it in the Synthesis-Tool of Mentor Graphics. But I had some problems with the Altera delay-Cells in the edif-out-netlist. Is there a apropriate way to "translate" the altera-design to xilinx. I heard that there is a way with Exemplar but we do not have Exemplar. Anton Scherer Microswiss-Zentrum Nord-Sued SwitzerlandArticle: 3738
Hi, I gathered for Information about Actel, and what suprising, I got all Info which I asked for in about 2 weeks! Simply go to www.actel.com an request for Databooks and Demo-CD's. The Books are very interesting and easy to read... By this I've no points to be dissatiefied. Hope, this will help you DominikArticle: 3739
Hi, By using Autologic II, we tried to synthesise a design to map it on a Xilinx 3064 FPGA. Because this family doesn't support XBLOX we have synthesise the design by setting the Macro mapping switch to NONE. However, Autologic II still generates XBLOX like :INC_16U_16U..., GT_16U_16U.... etc. The Xact PPR tool aborts with the message XBLOX not supported by the 3000 family. Is there anybody who can give us a very good hint to resolve this problems. Thanks to reading this news, Thousands thanks to reply. Marc & GilbertArticle: 3740
***************************************************************** 1st On-line Workshop on SOFT COMPUTING (WSC1) August 19 (Mon.) - 30 (Fri.), 1996 Special Session on EVOLUTIONARY ELECTRONICS ***************************************************************** Call for Papers and Invitation to take part in the Discussion ***************************************************************** As part of the 1st on-line workshop on soft computing, we call for papers for the special session on 'Evolutionary Electronics'. The session is intended as a forum for discussing the use of evolutionary computing techniques in electronics design and manufacturing, and the application of reconfigurable logic as the medium for the rapid prototyping of (evolved) digital logic designs. The session welcomes survey and tutorial style papers. The areas of interest include but are not limited to : a. Evolving anlogue circuits b. Intrinsic/extrinsic evolvable hardware c. VLSI placement, routing and testing using evolutionary algorithms d. Hardware implementations of evolutionary algorithms e. Evolutionary co-design f. Embryonics g. Implicit fault tolerance afforded by evolutionary design h. Field programmable gate arrays (i. Hardware implemented neural networks) The contributions can be in the form of : a) 6 page (max) Original paper, to be published in the workshop proceedings. In the case of survey or tutorial style papers authors are encouraged to submit the paper in several parts (such as part I and part II with the same title). b) 6 page (max) paper on your already published work c) YOUR PARTICIPATION in the discussion during the workshop PLEASE NOTE: The DEADLINE FOR SUBMISSION of abstracts and main text is Aug. 5, 1996 WSC1 will accept both ORIGINAL and ALREADY PUBLISHED papers (in electronic form). It is intended that original papers will be published in the conference proceedings. Please refer to the ORIGINAL CALL FOR PAPERS below for the submission and further details. To submit a paper for the special session, use the same submission procedure as for the WSC1 workshop, but, please, CLEARLY INDICATE THAT YOUR PAPER IS FOR THE SPECIAL SESSION ON EVOLUTIONARY ELECTRONICS. For any queries about the special session (only) please send email to: a.j.hirst@open.ac.uk http://kmi.open.ac.uk/~monty/evolelec.html To contact the WSC1 organisers: http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/ ============================================================== ORIGINAL CALL FOR PAPERS =============================================================== This is the preliminary CFP for the 1st Online Workshop on Soft Computing. This is also available on WWW. The URL is http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/ CALL FOR PAPERS The 1st Online Workshop on Soft Computing (WSC1) Aug. 19 (Mon) - Aug. 30 (Fri) On Internet (WWW (World Wide Web) ) Served by Nagoya University We call for your papers for online discussions using the Internet. The purpose of this workshop is to give its attendees opportunities to exchange informaiton and ideas on various aspects of Soft Computing and "save travel expenses" without having to visit foreign countries. We aim to have ample discussion time between the authors and attendees make them visible to everyone on the Internet. This is the 3rd workshop of On-line Workshop series on Internet. The first on-line workshop(WEC1) was held on Oct. 9 - 20, 1995. In spite of the very short preparation time ( 2 weeks ) for the workshop, 22 papers were submitted and great many people visited the home page: http://www.bioele.nuee nagoya-u.ac.jp/wec Encouraged by the unprecedented repercussions, we held the second on-line workshop(WEC2) on March 4 - 22, 1996. More people showed great interest into the workshop. You can still visit the home page at: http://www.bioele.nuee.nagoya-u.ac.jp/wec2 The scope of the 3rd on-line workshop is expanded from those of the preceding workshops. We call for your papers on Soft Computing. The great merit of this workshop series is that you can save travel expenses without having to visit and attendees and make them visible to everyone on the Internet. The WSC1 will welcome original contributions. The proceedings is planned to be published. We all look forward to your active participation. TOPICS: Fuzzy Logic Neuro - Computing Genetic Computing Probabilistic Reasoning Chaos Computing Hybrid Systems IMPORANT DATES: Deadline for Submission of Abstracts and Main Text: Aug. 5, 1996 Opening the papers to the public: Aug. 12, 1996 Workshop Weeks: Aug. 19-30, 1996 SUBMISSION PROCEDURE: Send an abstract and main text to the address below. mail address is : wsc@bioele.nuee.nagoya-u.ac.jp The abstract should be in text file. The main text should be in text file or in ps ( PostScript ) file. The size of the paper should be either A4 ( metric ) or letter size. Papers should be written in Times or similar font style, 10 points or larger with 2.5 cm margins on all four sides. If you write your paper with LaTeX, you can use "IEEEtran.sty" and "wscsample.tex" which is available from http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/form/IEEEtran.sty http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/form/wscsample.tex Another alternative is that if you prepare a WWW page for your main text, please send us only the abstract and tell us the Internet address ( URL ), and yours will be linked to our WWW page. The steering committee will make your abstracts visible on the Internet. The home page address is: URL: http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/ *Note: This workshop will accept papers both original and already published. If your paper is not original, please clarify the source of your paper. Your paper will not be included in the proceedings. If your paper is an original one, you need to send ps file or camera - ready manuscript of your main text. The paper should not exceed 6 pages in length. Please also send us the copyright form by postal mail or facsimile. The proceedings including your original contribution will be published. The copyright form is available also from http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/form/copyright.ps http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/form/copyright.txt DISCUSSION PROCEDURE: 1. Read the abstracts. 2. Copy the main texts ( ps files ) of interested papers. 3. Send questions and comments to wsc@bioele.nuee.nagoya-u.ac.jp (The steering committee will send the questions to the authors and receive answers from the authors, and make the Q&A visible on the Internet.) 4. Read the answers from the authors on http://www.bioele.nuee.nagoya-u.ac.jp/wsc1/ 5. Repeat the above steps 3 and 4 until you are satisfied. OFFICIAL LANGUAGE: English (Notice should be made that the main texts in ps files should not include any fonts other than English.) For further information, contact: Takeshi Furuhashi Dept. of Information Electronics Nagoya University, Furo-cho, Chikusaku Nagoya 464-01, Japan Tel. +81-52-789-2792 Fax. +81-52-789-3166 Paper Submission Form Title : Authors : Affiliation : Address : Phone : Fax : e-mail : Source(if your paper is already published): Keywords : Link(if you have a WWW home page of the paper) : Abstract :Article: 3741
THE COMPANY Established in 1981, Mentor Graphics Corporation (NASDAQ:MENT) designs, manufactures, markets and distributes electronic design automation (EDA) software and provides professional services supporting its customers' complete design environments. The company is a leader in worldwide EDA sales, with revenues of $440,714,000 over the last reported 12 months. Mentor Graphics is the first EDA vendor to win the STAR (Software Technical Assistance Recognition) award, and the only EDA vendor to win the award twice. The award is given annually by the Software Support Professionals Association (SSPA) for service excellence. The company currently employs approximately 2,400 people worldwide. In addition to its corporate offices, Mentor Graphics has sales, support, software development and professional services offices worldwide. The company's headquarters are located at 8005 S.W. Boeckman Road, Wilsonville, Oregon 97070-7777. World Wide Web site: http://www.mentorg.com. TECHNICAL MARKETING ENGINEER Design Creation Products DUTIES/RESPONSIBILITIES Provide worldwide technical field support for MGC's new line of Windows-PC based design entry products. Develop and deliver product demonstrations, support customer visits, gather customer requirements, work with engineering to design/enhance products, develop and deliver quarterly AE training, support tradeshows, workshops and seminars. Develop and deliver technical papers, applications notes and technical articles. QUALIFICATIONS: EDUCATION, EXPERIENCE, SKILLS, ETC. The ideal candidate will have: a BSEE with 2-3 years experience with Windows based PC EDA applications;2-3 years of Top Down hardware design experience with FPGA or ASICs using schematic and HDL based design techniques; Excellent oral and written skills; Experience with Client-Server applications using UNIX-Windows products a plus. 1-2 years of field applications experience highly desirable. Contact: John Huber, 503/685-1372, john_huber@mentorg.comArticle: 3742
In article <DuyvHJ.50J@news.uwindsor.ca>, hajimow@engn.uwindsor.ca wrote: > I am using Xilinx 4000 series and I want to design a dual port RAM. Pelase don't even try to do that. The XC4000E has additional hardware ( transistors ) that facilitate dual-port RAM operation. The XC4000 does not. If you need true dual port RAM operation, use the XC4000E. As a bonus, it is also cheaper than the equivalent XC4000. Yes, you can fake a dual-port RAM in XC4000 by time-division multiplexing a RAM, but that awkward method is now meaningless since the XC4000E is available. Peter Alfke, Xilinx ApplicationsArticle: 3743
In article <31F4FA06.C68@crpht.lu> Wilwert Marc <marc.wilwert@crpht.lu> writes: >Hi, > > By using Autologic II, we tried to synthesise a design to map it on a >Xilinx 3064 FPGA. Because this family doesn't support XBLOX we have >synthesise the design by setting the Macro mapping switch to NONE. >However, Autologic II still generates XBLOX like :INC_16U_16U..., >GT_16U_16U.... etc. >The Xact PPR tool aborts with the message XBLOX not supported by the 3000 >family. >Is there anybody who can give us a very good hint to resolve this >problems. If this is a new design you might want to consider an XC3064A or XC3164A, which I think (but am not sure) is supported by XBLOX. Good luck, PhilipArticle: 3744
In article <DuyvHJ.50J@news.uwindsor.ca> hajimow@engn.uwindsor.ca writes: > I am using Xilinx 4000 series and I want to design a dual port RAM. I >have got an application note that shows how to design it with 4000E series >that CLBs are edge triggered.Now my question is how should I change that >design in order to use 4000 series instead.Basically If I be able to make >RM 16x1S that is available in 4000E libraries by RM 16x1 that I have in >4000 libraries now, my problem is solved. Any comment is highly >appreciated. >Sayed-Hossain Hajimowlana Dept. Of Elec. Eng. The fully synchronous dual port RAM can not be emulated with the RAM in the XC4000. Dual Port is a primitive, like an AND gate. i.e. if all I have is AND gates, how do I emulate an inverter? If you are running your system at a reasonably slow clock rate (10 MHz or less), then you could build a memory that supported multiple ports by time multiplexing them with a faster clock within 1 cycle of your system clock cycle. You will still have to face the challenges of creating a valid write pulse within a stable window of address and data. You need to read the application data in the data book, on page 8-127 thru 8-147. good luck Philip Freidin.Article: 3745
I am just starting to look into VHDL, and I am having a hard time locating any kind of evaluation board/software with a Xilinx fpga, and some inputs/outputs. Does anyone know where I can buy one from? I don't really care what family of Xilinx, but I would prefer something in the 4000 line. -- +----------------------------------+ | mailto:bharstad@minn.net, | | Software Geek | | Mispellers of the world UNTIE! | | | |Looking for something to do in MN?| | http://www1.minn.net/~bharstad | +----------------------------------+Article: 3746
In article <Duv0nH.F9M@news.uwindsor.ca>, safiri@engn.uwindsor.ca wrote: I was wondering whether there is any design that > uses less resources of FPGA than 8-bit adder/subtractors to increment an 8-bit number by one. If you run at high speed (10 to 100 MHz ) or if you use only a few of these incre/decrementers, then the XC4000 or XC4000E will give you the most economical solution ( 2 bits per CLB ). If you are slow and also have many of these functions in the same chip, you could use serial arithmetic with the data stored in the small RAMs, and you would only need 2 CLBs per 16- or even 32-bit incre/decrementer ( plus a few common overhead CLBs ). There are some scientific instrumentation applications that thus can take advantage of distributed RAM, but it means that you need n clock ticks to operate on an n-bit word. Cute, but too slow for most applications. Peter Alfke, Xilinx ApplicationsArticle: 3747
Hi; Does Xact (ver5.2 support 4000E series? I didn't find it in its library. ThanksArticle: 3748
Hajimowlana Sayed wrote: > > Hi; > > Does Xact (ver5.2 support 4000E series? I didn't find it in its library. > > Thanks I think you need XACT 6.0.1 for PC's or 5.2.1 for Unix to support 4000E. Regards, ScottArticle: 3749
<jonathan@dcs.gla.ac.uk> wrote: >On 11 Jul 1996, GAMRAT Christian wrote: >> Is There anybody out there who knows anything new about the Xilinx XC6200 >> partially reconfigurable fpga ? >depends what you mean by new? i don't think there's any more information >available yet, preliminary timing details that i'm probably not allowed to >disclose even if i had the timing diagram where i could find it ;-) Xilinx have a 30-odd page data sheet on the XC6200 on their web site, with all the advance timing information included. The latest one is dated June 1st, 1996. I could not find any information in the Xilinx databook to support the assertion that the data in a flip-flop is cleared if the cell that contains it is being reconfigured. Certainly the Atmel devices allows a cell's logic to be modified while retaining the state of the flip-flop in the cell. I would be interested if anyone could clarify the point that the XC6200 would clear the register state when the remainder of the CLB was being modified. I also noticed that IBM are now entering the FPGA market, with a new architecture, based closely on the Atmel 6000 series (formerly the Concurrent Logic devices). The Atmel cell as been modified, improving routing and cell functionality, but retaining its interesting features such as partial and dynamic reconfigurability. Gordon Communications Division Email: g.mcgregor@eee.strath.ac.uk Electronic & Electrical Eng. Dept. University of Strathclyde Tel: +44 (0)141 552 4400 x 2250 Glasgow G1 1XW Fax: +44 (0)141 552 4968
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