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
In article <3701F050.DCF70028@irisa.fr>, Steven Derrien <sderrien@irisa.fr> wrote: >I was wondering if such an operation requires the download of a whole >frame to the Virtex ? It it is the case, this leads to a significant >reconfiguration overhead , (a few hundred microseconds ) Am I right? > It's much faster than that. Each Virtex device has a different frame size with the popular V300 device having a frame size of 672 bits. Loading just the frame data in the SelectMAP mode (8-bits) @ 50MHz would only take 1.68uS (672/8*20ns). EdArticle: 15576
Thanks for the Rockwell link. Have you looked at the Maxim 3885? 2.5 GBps 1-16 demux, Single 3.3V supply, LVDS out, only 660 mW. Price is right, too! http://209.1.238.246/1st_pages/1905.htm (via) http://www.maxim-ic.com/Datasheets.htm One of our major challenges in using this part has been coming up with a way to sychronize multiple 3885s without embedded framing information in the data. It's possible, but not pretty. NEL Japan has some (relatively expensive) demux parts good up to 10 GHz. (slow) http://www.nel.co.jp/uhe/index.html It looks like they are trying to get data sheets online, without much success so far. Here's the GIGA link. Only 622 MBps - too slow for me (same for the Fujitsu part). http://www.giga.dk/ As for the Altera LVDS speed question, the data sheet is silent on this issue but the press release claims 622 MBps, which is probably much higher than the useful system clock rate of the chip anyway. http://www.altera.com/html/products/apex.html regards, Tom Stephen King wrote: > > I have recently been looking at this problem and have come across the > following devices. Not having used them I can not speak as to their > suitability or availability. > > GIGA GD16333 4:32 demux > > Fijitsu FMM4005EK 2:16 demux > > Rockwell 5:64 demux ????? > www.rsc.rockwell.com/converters/tech_8-64.htm > > Hope that this is of some help. > > If you know of any other integrated solutions to this problem I would be > interested to know. In addition does any one know how fast the LVDS inputs > on the Altera 20KE are likely to work? > > Stephen King > CRL > sking@crl.co.uk > > Pierre Langlois wrote: > > > All the fast (>100 MHz) ADCs I know of have ECL outputs (SPT 7750, > > HI116, etc.). > > > > With FPGAs able to do process data at such high data rates, the > > difficulty now seems to get the data to the FPGA. Level translators are > > available on the market, but seriously complicate system timing > > especially if multiple chips must be used. > > > > Does anyone know of FPGAs with ECL-compatible inputs? Any plans to > > bring some to market in the near future? > > > > Alternatively, does anyone have any comments on their experience > > interfacing FPGAs to fast ADCs? > > > > Thanks in advance. > > > > =============================================== > > Pierre Langlois > > Département de mathématiques et informatique > > Collège militaire royal du Canada > > langlois-p@rmc.ca > > =============================================== -- Tom Burgess Digital Engineer National Research Council of Canada Herzberg Institute of Astrophysics Dominion Radio Astrophysical Observatory P.O. Box 248, Penticton, B.C. Canada V2A 6K3 Email: tom.burgess@hia.nrc.ca Office: (250) 490-4360 Switch Board: (250) 493-2277 Fax: (250) 493-7767Article: 15577
Daryl Bradley wrote: > I've almost got to the point of giving up trying to ask VCC for advice > before our lab purchase some of their virtex virtual workbench > development boards. Hopef ully someone here may be able to help me with > the following > > The Virtual workbench is a stand alone board - does anyone know if cable > to connect to the pc to download the configuration (sorry not too hot on > the terminology - that's why we want the boards - to learn!) are > supplied, or if not, why type needs to be purchased. > > I have also been advised that the foundation software should be one of > the most appropriate for us to use - any comments on this? > > I certainly hope someone can help me as the funding we have has to be > used very soon otherwise we lose it! > > Many thanks for any advice anyone can give > > Daryl Bradley The download cable must be purchased separately. The board supports any download cables designed for use with the Virtex. -- Steve Casselman, President Virtual Computer Corporation http://www.vcc.comArticle: 15578
Daryl Specter here from Insight Electronics, LAX, USA The product you are referring to will be available very shortly from both Virtual Computer Corp in Reseda, Ca, USA as a standalone product, or additionally from Insight Electronics in the USA, Canada, & Mexico as well as from MEMEC in Europe. Insight / MEMEC will be offering the board standalone as well as coupled with Tools (Alliance Series, if you already have a front-end tool like Synplify; and Foundation Series, if you need al the tools to capture and compile a design). that come with the necessary download cable to load the part. Daryl Specter ps: I also find it useful to have the X-Checker cable that Xilinx sells (Avail throught the same channels) as it is faster when downloading. The board also contains a flash memory that you can reporgram in your favorite programmer (stay tuned, one of the applications being looked at is the ability to reprogram the flash on board) and will be supported via a web page soon to be released to the public as a subpage of Insights WEB page (www.insight-electronics.com) Daryl Bradley wrote: > I've almost got to the point of giving up trying to ask VCC for advice > before our lab purchase some of their virtex virtual workbench > development boards. Hopef ully someone here may be able to help me with > the following > > The Virtual workbench is a stand alone board - does anyone know if cable > to connect to the pc to download the configuration (sorry not too hot on > the terminology - that's why we want the boards - to learn!) are > supplied, or if not, why type needs to be purchased. > > I have also been advised that the foundation software should be one of > the most appropriate for us to use - any comments on this? > > I certainly hope someone can help me as the funding we have has to be > used very soon otherwise we lose it! > > Many thanks for any advice anyone can give > > Daryl BradleyArticle: 15579
It seems that there are lot of people interested in the idea of free hardware cores just like me. I have lot of links on IP cores and free circuit project at my site at Links to IPCores http://www.geocities.com/SiliconValley/Pines/6639/ip/ free circuit projects http://www.geocities.com/SiliconValley/Pines/6639/freecir/free_circuit.html I also try to contact those people who like this idea in order to establish a project that can design both hardware cores and software tools. if you like to join me please send me an email so as to organize our selves with other people on the netArticle: 15580
On Mon, 29 Mar 1999, Mark Rogers wrote: > I suspect that ultimately the IP industry will evolve until it very closely > matches the software industry. A few power house companies supplying the > bulk of the IP, weekenders and college students cranking out shareware IP, > and then some huge open source products being developed over the web. > > Anyway, food for though... > > Mark Are there any sources of sharware IP now? This sounds like a good idea for retired people too. --alvinArticle: 15581
Evan Samuel wrote: > > I'm looking for an affordable package for schematic capture, VHDL systhesis > & FPGA systhesis. I work at home and do not have the resources of a large > company. If any one can help me identify any sources please reply to the > post or email: evansamuel@earthlink.net. Do you have preferences toward any particular FPGA vendor(s) ?Article: 15582
Richard Guerin wrote in message <3702C58A.25A4F025@home.com>... >Evan Samuel wrote: >> >> I'm looking for an affordable package for schematic capture, VHDL systhesis >> & FPGA systhesis. I work at home and do not have the resources of a large >> company. If any one can help me identify any sources please reply to the >> post or email: evansamuel@earthlink.net. > > >Do you have preferences toward any particular FPGA vendor(s) ? I have worked with Xilinx, and some with Altera. I, however, do not have a particular preference. I was hoping to obtain a schematic capture or VHDL software with portable outputs to the FPGA vendors synthesis tools.Article: 15583
Could someone please recommend a website or a book for a general introduction to Reconfigurable Computing? Thanks PaulArticle: 15584
Actel has the best "free" tools - if you like OTP parts. Evan Samuel wrote in message <7dtk20$iqd$1@fir.prod.itd.earthlink.net>... >I'm looking for an affordable package for schematic capture, VHDL systhesis >& FPGA systhesis. I work at home and do not have the resources of a large >company. If any one can help me identify any sources please reply to the >post or email: evansamuel@earthlink.net. > >Thanks, >Evan Samuel > > >Article: 15585
Mandeep Singh <msingh@ececs.uc.edu> wrote in message news:3702F3D3.1B4EC598@ececs.uc.edu... > Could someone please recommend a website or a book for a general > introduction to Reconfigurable Computing? www.vcc.comArticle: 15586
How many designs do you turn around in a year? If the answer is many for different clients, you will probably find you'll need to spring for something that is more mainstream than you would perhaps otherwise need to do. I maintain unrestricted licenses for viewlogic and synplicity, plus Altera and Xilinx tools among others. I find that using the vendor independent tools helps with the changing gears from project to project and customer to customer. I hate to think of the crossed control keys I'd be getting if I were using each vendor's capture packages too. If this is for a hobby or once in a while use, I wouldn't spend the $ on those tools. The programs packed in the vendor tools are adequate for occasional use, and far cheaper. Evan Samuel wrote: > I'm looking for an affordable package for schematic capture, VHDL systhesis > & FPGA systhesis. I work at home and do not have the resources of a large > company. If any one can help me identify any sources please reply to the > post or email: evansamuel@earthlink.net. > > Thanks, > Evan Samuel -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randrakaArticle: 15587
Ed McGettigan wrote: > In article <3701F050.DCF70028@irisa.fr>, > Steven Derrien <sderrien@irisa.fr> wrote: > >I was wondering if such an operation requires the download of a whole > >frame to the Virtex ? It it is the case, this leads to a significant > >reconfiguration overhead , (a few hundred microseconds ) Am I right? > > > > It's much faster than that. Each Virtex device has a different frame > size with the popular V300 device having a frame size of 672 bits. > Loading just the frame data in the SelectMAP mode (8-bits) @ 50MHz > would only take 1.68uS (672/8*20ns). > > Ed Does a frame corresponds to the global reconfiguration of a small area of the Virtex (CLB logic and Routing), or are there different types of frames ? For example some frames for routing configuration and some others for CLB/IOB ? According to the virtex datasheet, a XCV300 has a 32x48 CLB matrix, the whole device needs 1,700,000 bits to be reconfigured, This means that the "average" configuration data for a CLB plus routing is 1,100 bit i.e 2 frames. I I'm not wrong, this is much more interesting than I expected, as I understood that only a whole column of the Virtex could be reconfigured at a time ... By the way ,does someone can tell me about the availability of the JBit API ? (Xilinx Web site says Q1 1999) but there is not much informations about it. Thank you, Steven DerrienArticle: 15588
In article <Pine.BSI.3.96.990331115256.1213A-100000@malasada.lava.net>, "Alvin E. Toda" <aet@lava.net> wrote: > On Mon, 29 Mar 1999, Mark Rogers wrote: > > > I suspect that ultimately the IP industry will evolve until it very closely > > matches the software industry. A few power house companies supplying the > > bulk of the IP, weekenders and college students cranking out shareware IP, > > and then some huge open source products being developed over the web. > > > > > > Mark > > Are there any sources of sharware IP now? This sounds like a good idea > for retired people too. > > --alvin > There are some, but its rather scattered. The following sites have some (rather VHDL biased; does anyone know equivalent sites dealing mainly with Verilog?) http://www.vhdl.org/vi/fmf http://umunhum.stanford.edu/PAM-Blox http://erm1.u-strasbg.fr/db/ http://freecore.com http://tech-www.informatik.uni-hamburg.de/vhdl/ Is there enough interest to set up a mailing list on this topic? Graham -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your OwnArticle: 15589
I already have those papers, but they are no good. I am looking for a bitparallel implementation, not bitstream. All the articles they present are pretty much all the same :) Does anyone have experience with this ? Alex Rast wrote in message <37001f6e@news.nwlink.com>... >Yes. Check out: >M. Gschwind, V. Salapura, O. Maischiberger,"A Fast FPGA Implementation of a >General Purpose Neuron", FPL '94 Proceedings, Springer-Verlag 1994, >ISBN# 3-540-58419-6 > >And also see their paper "A Generic Building Block for Hopfield Neural >Networks with On-Chip Learning". > >Indeed, many of their papers, and plenty of other useful references, are at > >http://www.vlsivie.tuwien.ac.at/mike/custom-computing.html >Article: 15590
In article <370323D8.68D63B6F@irisa.fr>, >Does a frame corresponds to the global reconfiguration of a small area of >the Virtex (CLB logic and Routing), or are there different types of >frames ? For example some frames for routing configuration and some >others for CLB/IOB ? > A Virtex frame contains a sequence of bits that spans the entire height of the device with multiple frames per column. >According to the virtex datasheet, a XCV300 has a 32x48 CLB matrix, the >whole device needs 1,700,000 bits to be reconfigured, This means that the >"average" configuration data for a CLB plus routing is 1,100 bit i.e 2 >frames. > You're forgetting the IOB configuration and the Block SelectRAM configuration space as well as configuration overhead. >If I'm not wrong, this is much more interesting than I expected, as I >understood that only a whole column of the Virtex could be reconfigured >at a time ... To reconfigure all bits in a CLB column including CLB configuration, LUT equation and routing requires 48 frames. However you do not need to write all 48 frames and this can lead to some interesting uses. > >By the way ,does someone can tell me about the availability of the JBit >API ? >(Xilinx Web site says Q1 1999) but there is not much informations >about it. > I do not have information on the Virtex version of JBit (AKA JavaBits). But obviously we have missed the deadline for 1Q99. EdArticle: 15591
Mandeep Singh wrote: > > Could someone please recommend a website or a book for a general > introduction to Reconfigurable Computing? Barr, Michael. "A Reconfigurable Computing Primer" Multimedia Systems Design, September 1998, pp. 44-47. Also available online at: http://www.netrino.com/Articles/RCPrimer/ Cheers, MichaelArticle: 15592
Does any one want to talk about Dynamic Configuration? Here is my ICQ # : 33301195 Contact me. Riad BourguibaArticle: 15593
here is my ICQ # 33301195 Riad BourguibaArticle: 15594
At Last! A Truly Moderated Safe Email list! Get your message out to thousands without spamming! Lucrative two tier reseller option. http://dalesafe.com/index.cgi/AK1433 udhdusuvdtbptxhsdzzcmubswdvifclgmlgplrArticle: 15595
If you'd like to maintain some sort of vendor independence then the next obvious question is what do you consider affordable ? If you're a power user and have a power budget (approx $12K) .... IMHO, Mentor Graphic's Packaged Power suite (Renior, ModelSim, Exemplar Leonardo Spectrum) is a tough value to beat. http://www.packagedpower.com/mgc/ Alternatley, I agree with a previous post that Actel presently has the very best FREE tools ( VeriBest, Synplify-Lite, Designer-Lite ... until Jan 2000). Checkout: http://www.acteldesktop.com/ For other FPGA software links check out: http://www.optimagic.com/lowcost.shtmlArticle: 15596
Speaking of high speed serializer/deserializer interfacing to FPGAs... Can somebody tell me what the Fibre Channel folks were thinking when they came up with the standard that said that all these little PQFP-64 gigabit serializer/deserializer chips (e.g., AMCC S2070 and the bazillion clones thereof) should have 2 ~50MHz clock outputs (with 10 bit encoded parallel outputs) instead of 1 ~100MHz clock output? Given that the same 8B/10B encoding is used with gigabit Ethernet, all the SerDes vendors optimized their chips for it's 125MHz clock rate and kept the same interface to boot. Grrrr... We're going through some effort at the moment to get back to one 100MHz byte stream. Since the transmitter side of the SerDes chips want 1 100MHz clock with a bytewide input, my only guess is that the Fibre Channel guys thought you'd design something along the lines of: -- 50MHz dual (parallel) 8B/10B encode of 16 bit data followed by a simple mux running at 100MHz. Since the encoding consumes a significant amount of logic, presumably it was hard at some point to build a 100MHz encoder but not so hard to build a 100MHz mux. -- 50MHz dual 10B/8B decode, and we're just going to assume your system interface is 16 or 32 bits wide anyway so there's no need to get a 100MHz clock back. The deserializers go to some effort to guarantee that commas show up on a particular clock phase, so I'm tempted to think my guess here is a good one. ---Joel KolstadArticle: 15597
What about it? I did a paper in '96 on a video processor that used dynamic configuration for the frame processing. You can get a copy of that paper from my web site. Bourguiba Riad wrote: > Does any one want to talk about Dynamic Configuration? > > Here is my ICQ # : 33301195 > > Contact me. > > Riad Bourguiba -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randrakaArticle: 15598
Are you sure about that price? Seems pretty low to me. Richard Guerin wrote: > If you'd like to maintain some sort of vendor independence then the next > obvious question is what do you consider affordable ? > > If you're a power user and have a power budget (approx $12K) .... > IMHO, Mentor Graphic's Packaged Power suite (Renior, ModelSim, Exemplar > Leonardo Spectrum) is a tough value to beat. > > http://www.packagedpower.com/mgc/ -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://users.ids.net/~randrakaArticle: 15599
Hello, I have taken up a project to build a board as a CDMA baseband receiver. Basically I am planning to use an Altera FLEX10K device to implement it. A question I encountered is that how to build a Digital Matched Filter in FPGA. (Maybe about 256 taps or more? Is this parameter too difficult to realize?) I think the big barrier is that how to realize the high-speed parallel adder. (The chip rate is about 4M cps. The sample rate for acquisition is 4M sps too.) Could anyone give me some hints on this? Any help is greatly appreciated. :) Thanks in advance. Forest Niuman niuman@263.net -----------== Posted via Deja News, The Discussion Network ==---------- http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
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