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 h0h@news.tv.tek.com, bobe@prebman.tv.tek.com (Bob Elkind) writes: > davelec@extro.ucc.su.OZ.AU (David le Comte) writes: > >kenn@neocad.com (Kenn Perry) writes: > > > >> ... PC prices begin > >>at $995 U.S. list. NeoCAD's new, low-cost Access system > >>provides support for all vendor's devices under 3K gates ************** > >>(approximate), timing-driven place and route, an EDA vendor > >>integration kit for $2,995. This configuration has been very > | > I *think* you mean $12,995. -- be > > >>attractive to new users of FPGAs. I think the $2,995 is a *new* deal from NeoCAD. Notice the *under 3K gates*. Sounds smart to me. This is a cheap way to get vendor independence and try out NeoCAD. The other $4,995 single vendor deals claim to give you a "better" router with a *path* to vendor independence. You might try getting a new quote if less than 3K gates does anything for you. --- - Bill Wolf, Raleigh NC - My opinions, NOT my employer'sArticle: 651
The Cypress Warp2 VHDL compiler supports 22v10 devices. The entity code for 24-pin devices exist, however there is no entity code for 28-pin devices. - Does anybody know if it is possible to use 28-pin 22v10 in Warp2? - Does anybody have any idea for fitting designs into 28-pin 22v10 in Warp2? I would appreciate any help. -- Akinori Sugiura Email: sugiura@tcp-ip.or.jp Tel & Fax: +81 532 25 8374Article: 652
In article <D31C10.2tM@world.std.com>, Bryan Butler <butler@world.std.com> wrote: >I am having problems programming an in-circuit Intel/Altera FX780 device. >I get frequent verify errors and occasional "unable to open device" messages. >Also, it seems like once I program it, I have to cycle the power to program it >again. (I'm only trying to program the SRAM at this point, not the PROM). I've never seen this happen before. The command for programming the RAM goes over the JTAG port and should not need a power-on reset to be effective. > >I'm using the Intel flex cable and programming software (pengn). Although I >lengthened the cable for convenience, this doesn't seem to have any negative >effect. If anything, it seems more reliable with the longer cable! The only time I've seen this is with noise on the TCK signal. You might try a 7414 to clean up the clock at the 780 end of the cable. How did you lengthen the Intel cable? > >Anyone else notice problems with these devices? > >-- >------- >Bryan Butler >butler@world.std.com -- || Dave Van den Bout || || Xess Corporation ||Article: 653
Hello, i'm looking for following synthesis tools for the Linux platform. I believe most of them if not all are University tools. Thanks much. bdsyn espresso sis/misII -gshin PS BTW, i'm in search of tool like bdsyn but can directly work on FSM rather than just on the combinatoric logic portion.Article: 654
Hi, I am looking for someone to share a room during the FPGA Symposium in Monterey from 12-14 Feb. If you are looking for a room-mate please contact me by email. Thanks, Shashidhar -- -- E-mail: thakur@cs.utexas.edu Phone: (512) 471-9753 Office: PAI 5.38 World-Wide-Web: http://www.cs.utexas.edu/~thakur/Article: 655
Hi, I was wondering how people are dealing with the following situation or similar situations. I am using Viewlogic's Prosynthesis to target an XC4000 Xilinx FPGA from a VHDL-based design. The VHDL has been written, the logic gates have been synthesized, and the resulting logic has been simulated. The simulation looks good, and I am now about to use the XACT 5.0 tools to go through the PPR process. However, since the synthesis tools, only use basic parts of the XC4000 library (i.e. AND, NAND, NOR, OR gates and FDCE (flip flops), etc), the FPGA targetting appears extremely inefficient. I created a "top-level" which consisted of hierarchical elements which totalled about 4000 "cells" (gates and flip flops). I used the "estimate" option on the PPR, and I targeted a 4013 FPGA. The estimate was that 95% of the function generators were needed to be used. My understanding is that the only way around this poor use of resources is to instantiate the library parts in the VHDL code. That is, XC4000 adders, multiplexers, etc. that would map to the FPGA more efficently could be explicitly used. This seems to take a away from the whole point of the VHDL design in the first place! I would have just used a schematic capture approach instead! Please tell me if there are any solutions to this dilemma. Is this inefficient use of FPGA resources just a consequence of a high-level design approach (i.e. VHDL) that must be endured? Thanks, EdArticle: 656
In article <3gb0g1$68h@aurns1.aur.alcatel.com> wolf@aur.alcatel.com writes: >In article h0h@news.tv.tek.com, bobe@prebman.tv.tek.com (Bob Elkind) writes: >> davelec@extro.ucc.su.OZ.AU (David le Comte) writes: >> >kenn@neocad.com (Kenn Perry) writes: >> > >> >> ... PC prices begin >> >>at $995 U.S. list. NeoCAD's new, low-cost Access system >> >>provides support for all vendor's devices under 3K gates > ************** > >> >>(approximate), timing-driven place and route, an EDA vendor >> >>integration kit for $2,995. This configuration has been very >> | >> I *think* you mean $12,995. -- be >> >> >>attractive to new users of FPGAs. > >I think the $2,995 is a *new* deal from NeoCAD. Notice the *under 3K gates*. >Sounds smart to me. This is a cheap way to get vendor independence and >try out NeoCAD. The other $4,995 single vendor deals claim to give you a >"better" router with a *path* to vendor independence. You might try getting >a new quote if less than 3K gates does anything for you. > >- Bill Wolf, Raleigh NC >- My opinions, NOT my employer's You are right. Kenn Perry from Neocad replied via email to clear up my confusion. To sum up: Neocad has three product lines -- Access -- under 3K gates, entry level product, starting at $995 US Catalyst -- up to (and including) 15K gates Paragon -- includes >15K gate devices If there are differences in the place/route tools, or just differences in the supported device libraries (for the different size ranges of devices), I'm not sure. There are options within the Catalyst and Paragon lines for single/dual/all vendor/FPGA/lib support (e.g. Xilinx 3K, ATT 2C, Actel, Moto, etc.) There are upgrade paths from the lower product lines to the higher product lines. Talk to a NeoCad rep (or kenn@neocad.com) for specific prices. There! I've told you all that I know (and perhaps a bit more!). I don't work for NeoCad, I'm just a customer. Bob Elkind, Tektronix TV ProductsArticle: 657
Can anyone tell me how or where to obtain specifications for the Library of Paramterized Modules (LPM)? Thanks for your help. Matt Henry Sandia National LaboratoriesArticle: 658
Mark, Your options are actually as follows: Synthesis: Implementation: Exemplar/ \ XACT/ MINC (with VHDL option)/ -----> NeoCAD ABEL (with VHDL option)/ / another synthesis tool The synthesis options in the first column (amongst many others) will each provide an output suitable for implementation with Xilinx XACT or NeoCAD's FPGA Foundry (these are the only two implementation tools on the market for Xilinx). The means of conveying the data from the synthesis tool to the implementation tool is usually either an EDIF or XNF netlist. I won't comment on the merits of the tools mentioned, since I'm not impartial (my employer is UK VAR for two or three of the products you mention)! You will only get back-annotated design files for simulation out of the implementation tools (XACT and NeoCAD) since Xilinx performance is non-deterministic before place-and-route. For the same reason, timing simulation is a worthwhile activity. Sorry, we don't have a FAQ for this group (yet), the nearest FAQ is in comp.lang.vhdl, which lists all available synthesis tools. Hope this helps. BTW, don't worry, you *do* get a download cable with NeoCAD's FPGA Foundry! David Pashley < ------------------------ < < < ---------- Email: david@fpga.demon.co.uk | Direct Insight Ltd < < < < > Tel: +44 1280 700262 | | *The EDA Source* < < < Fax: +44 1280 705196 | --------------------------- < ------------------------------------------ In article <3g8835$59o@doc.armltd.co.uk> msnook@armltd.co.uk writes: "I currently have a number of behavioral VHDL models that I need to target "into Xilinx FPGAs. I currently use the Model Tech VHDL simulator and am "evaluating the best route to a routed Xilinx 3000 series part. I am not "interested in automatic partioning or other vendors FPGAs at this stage. I "am Sun Sparc based so price comparisons with PCs are not really valid. " "As I see it I have four options: " "1> NeoCAD "2> Exemplar coupled with XACT "3> MINC VHDL input filter and XACT "4> ABEL VHDL input filter and XACT "....<snip>Article: 659
In article MsB@indirect.com, seeker@indirect.com (Stan Eker) writes: >Ken Yiu (kky@itd.dsto.gov.au) wrote: >: Hi. >: This is not an appropriate forum for this question, >: so my apologies in advance. I am, however, not aware >: of any appropriate news group. > >None of the normal groups want to encourage suppliers and vendors to trash >the 'net. Your normal recourse should be through alternate distribution, >and especially NOT through this group. Try calling the factory, first, as >they'll have a better handle on who has what parts. > >If you don't think distributors & vendors are slavering at the thought of >forcing their vapid advertising in the faces of 10 million possible >customers, then you aren't very familiar with the breed. Most salesmen have >all the tact and morals of a weasel, for the same reasons. > I appreciate your views - and share them to an extent. I have tried Xilinx direct, and the Xilinx worldwide distributors network - in fact this post was a last resort/act of desperation. FYI, it seem that independant component distributors (e.g. America II etc) are the most useful group in this type of situation. The factory is NOT useful, as they always quote you 2 month lead times. The distributor is more likely to be able to source (limited) stocks. I apologise if I offended/bothered you with the post. Regards, Ken. Disclaimer: These views are not necessarily those of my employer.Article: 660
George Shin (gshin@netcom.com) wrote: : Hello, i'm looking for following synthesis tools for the Linux platform. I : believe most of them if not all are University tools. Thanks much. : bdsyn : espresso : sis/misII : -gshin : PS BTW, i'm in search of tool like bdsyn but can directly work on FSM rather : than just on the combinatoric logic portion. I compiled espresso on my Linux PC more than a year ago and I think it compiled straight out the box without any modifications. Try archie to search for espresso. -- -------------------------------------------------------------------- Petter Gustad E-mail: pegu@dolphinICS.no Hardware Design Engineer Tel.: +47 22 62 70 00 Dolphin Interconnect Solutions A.S. Fax.: +47 22 62 71 80 Street address: Mailing address: Olaf Helsets vei 6 P.O. Box 70, Bogerud Oslo, Norway N-0621 Oslo, Norway --------------------------------------------------------------------Article: 661
> >Hi, > > I was wondering how people are dealing with the following situation or >similar situations. > > I am using Viewlogic's Prosynthesis to target an XC4000 Xilinx FPGA >from a VHDL-based design. The VHDL has been written, the logic gates >have been synthesized, and the resulting logic has been simulated. >The simulation looks good, and I am now about to use the XACT 5.0 tools >to go through the PPR process. > > However, since the synthesis tools, only use basic parts of the >XC4000 library (i.e. AND, NAND, NOR, OR gates and FDCE (flip flops), etc), >the FPGA targetting appears extremely inefficient. I created a >"top-level" which consisted of hierarchical elements which totalled about >4000 "cells" (gates and flip flops). I used the "estimate" option on >the PPR, and I targeted a 4013 FPGA. The estimate was that >95% of the function generators were needed to be used. > > My understanding is that the only way around this poor use of resources >is to instantiate the library parts in the VHDL code. That is, >XC4000 adders, multiplexers, etc. that would map to the FPGA more >efficently could be explicitly used. This seems to >take a away from the whole point of the VHDL design in the first place! >I would have just used a schematic capture approach instead! > > Please tell me if there are any solutions to this dilemma. Is this >inefficient use of FPGA resources just a consequence of a high-level >design approach (i.e. VHDL) that must be endured? > >Thanks, > > Ed > The _SOLUTION_ to this problem requires the synthesis vendors to SYNTHESIZE to higher level constructs. Then when they TARGET a particular device they can take advantage of the special features within that device. You might be able to get Exemplar or Synopsys to benchmark your VHDL code as both of them utilize SOME but not ALL (yet) of the Xilinx special functions. A compromise is to instantiate only the largest of the inefficient blocks and let the synthesis tool generate the others. Typically, counters and adders are the only components that need to be instantiated. Actually you will still receive significant benefits from using HDLs in the random logic and state machine areas. Be sure to complain loudly to the synthesis vendors or they will not move their intermediate synthesis form to a higher level representation! Good Luck, Charles SHELOR ENGINEERING VHDL Training, Consulting, and models 3308 Hollow Creek Rd (817) 467-9367 Arlington, TX 76017-5346 cfshelor@acm.orgArticle: 662
Article: 663
I am interested in experiences with FPGA designs in applications that demand >1 watt per package, specifically: 1. Packaging/cooling Do you have access to the package/cooling you need/want? 2. AC specs vs. junction temp: Did you have to derate AC specs because of elevated temp operation, did you even check the temp vs. speed specs, or did you cross your fingers? 3. What clock frequencies are you using? Were you limited in gate utilisation and/or operating frequency by power and/or cooling considerations? 4. Did you have unforseen problems? Surprises? etc. Or did your design tools give you some warning/hint that there might be timing vs. power/heat vs. utilisation problems? 5. If you had probs, what did you do (specifically) to address them to complete your project, and what are your plans for avoiding similar probs in the future? Do you see FPGA vendors responding to your past problems and/or future aspirations? 5a. (Essay Question [optional, extra credit!]) What are your convictions about the *technical* solutions to your specific or general problems? Be as specific or general as you like. As this list of questions probably intimates, we *have* had problems with heat vs. pkg. vs. timing; my agenda is to do some of the technical marketing research (across the net) that I think the FPGA vendors need/should do themselves. Depending upon the responses, I may find myself alone in an isolated FPGA applications backwater. We'll see. Naturally, the results will be posted to this newsgroup, edited to avoid commercialism, protect confidentiality, etc. Thank you for taking the time to respond. I have no financial interest in *any* FPGA or FPGA/EECAD vendor. Bob Elkind, Tektronix TV Test & Measurement bobe@tv.tv.tek.com 503.627.4417Article: 664
In article rad@darum.uni-mannheim.de, kugel@mp-sun6.informatik.uni-mannheim.de (Andreas Kugel) writes: > We want to use the JTAG port of the XC4000 familiy to configure these > devices. Unfortunately there is no detailed description on this topic > available. First, I suggest reading the XAPP note in the Xilinx 1994 Data Book (pp 8-37) It has quite a few details on data format, etc. > In detail my questions are: > how long must I wait from the UPDATE-IR state (instruction > configure) to the SCAN-DR state ? May I go there directly or > only thru the IDLE state ? I personally don't have any experience with this as we us a uP to drive the Jtag and always spend quite a bit of time in the Idle state. But I believe from discussion with Xilinx engineers, that passing through Idle is not required. > > If in SHIFT-DR state, may I load all config bits in one sequence > or must I go thru UPDATE-DR after every bit / frame / something ? The entire bitstream (as defined in the XAPP Note mentioned above) can be loaded as one Shift-DR. We do it regularly. > > If there is a DSDL file floating around ? (copy please) Sorry, I confess my ignorance, what is a DSDL. Be glad to answer any other questions (within myknowledge, of course) Barton Quayle Barton@qcktrn.com Quickturn Design Systems, IncArticle: 665
mrhenry@sandia.gov (Matthew R Henry,2274,4-2630) writes: : Can anyone tell me how or where to obtain specifications for the : Library of Paramterized Modules (LPM)? Thanks for your help. Ask for EIA/IS-103 "EDIF Library of Paramterized Modules (LPM)" published by: Electronic Industries Association Standard Sales Department (Attn: Cecelia Fleming) 2001 Pennsylvania Avenue, N.W. Washington D.C. 20006, USA Available from: Global Engineering Documents 1-800-854-7179 or 1-303-792-2181Article: 666
Hello ! I wonder if anyone has any information about: - commercially available "on-fly" reprogrammable devices, i.e. devices which can be programmed during their operation (or better part of the chip can be reprogrammed while another parts are still operational). I am interested in any type of device FPGA, PLD etc. (I am only aware of AT6000 family from Atmel) - any research activities on-going in this area Thanks in advance for any references or comments on this. Milan. o--------------------------------o-------------------------------o | Milan VASILKO (Mr.) | School of Electronics | | | Bournemouth University | | mvasilko@bournemouth.ac.uk | Fern Barrow | | tel: +44-(0)202-595 101 | Poole, Dorset BH12 5BB | | fax: +44-(0)202-595 314 | UNITED KINGDOM | o--------------------------------o-------------------------------oArticle: 667
This is our second posting to the internet of interesting design automation sources throughout the world-wide web. There was a very large volume of newsgroup readers who visited our "DA-related Information on the WEB", that we decided to continue offering this free public service. We also received additional pointers from our readers. If you have any additional favorites of your own, send them to webmaster@cadmazing.com and we will include them in next month's update. We will try to update this monthly (as time allows) so check back next month for intersting pointers. Here are the updates for Feb 95. To access these, set your url to http://www.cadmazing.com/cadmazing Power Modeling Electronic Design Automation Brief of Grand Unification Theory Grand Unification Theory List of EDA- and Design-Related Software Microelectronics and Computer Technology Corporation Multichip Systems Design Advisor (MSDA) Ptolemy VHDL Tid-Bits NanoFab Home Page ASIC & EDA September 1993 Issue (*demo*) SRC University Fabrication Facilities Database Directory of ISDATA Microelectronic Systems Newsletter Universities Listing of the U.C. Berkeley Design Technology Warehouse News groups Engineering-related newsgroups Newsgroup: comp.lsi.cad Newsgroup: comp.cad.cadence Newsgroup: comp.cad.compass Newsgroup: comp.cad.synthesis Newsgroup: comp.lang.verilog Newnews: Sci.electronics.cad by date Newnews: Comp.arch.fpga by date ESNUG (Synopsys user group) Archive by dateArticle: 668
Please unsubscribe me from the mailing list. Thanks.Article: 669
In article <22093.9502011123@bmth.ac.uk>, Milan Vasilko <mvasilko@bournemouth.ac.uk> wrote: ->- commercially available "on-fly" reprogrammable devices, i.e. devices -> which can be programmed during their operation (or better part of the -> chip can be reprogrammed while another parts are still operational). -> I am interested in any type of device FPGA, PLD etc. The Altera FLEXlogic 8160 has this feature in a more limited sense than the Atmel part: half of the 8160 can be reprogrammed while the other half is running. -- || Dave Van den Bout || || Xess Corporation ||Article: 670
I have an emacs mode for verilog.. I *thought* there was one out there for VHDL. Anyone seen or heard of one for Altera's HDL (AHDL) thanks in advance ! -tim schneider@email.iac.honeywell.comArticle: 671
In article <22093.9502011123@bmth.ac.uk>, Milan Vasilko <mvasilko@bournemouth.ac.uk> wrote: >I wonder if anyone has any information about: > >- commercially available "on-fly" reprogrammable devices, i.e. devices > which can be programmed during their operation (or better part of the > chip can be reprogrammed while another parts are still operational). > I am interested in any type of device FPGA, PLD etc. > (I am only aware of AT6000 family from Atmel) > Look into the CLAy family from Nat'l Semiconductor- they claim partial reconfigurability at the 'cell' level. A CLAy cell can implement functions of several gates and an optional flip-flop, so it's fine-grain reconfigurability. My only experience with these devices is a glance at their preliminary data sheet, so I can't advise on their practical utility. For more info, contact Tim Garverick, garv@berlioz.nsc.com. The usual disclaimer applies. -Aaron FerrucciArticle: 672
Hello, does anyone tried or have ported OCTTOOLS to Linux or have the patches to build them with gcc? Currently i have ported 'mustang', 'espresso' and got the 'sis' (successor to earlier 'misII'). Any help would be greatly appreciated. Thanks much... -gshin PS Is OCTTOOLS even available as PD or $$$ needed to Berkeley? PSS I have the Linux ported Ptolemey and was wondering if Ptolemey is part of OCTTOOLS package?Article: 673
Check out Lattice Semiconductor ISPLSI, a starter kit that includes window's software to compile logic blocks, In-Circuit Programmable/ very easy to use, the starter kit only contains ability to use 1016 devices which has 32 i/o pins and some input/clk pins as well and is $99.00. The upgrade is more expensive, I picked it up for $650.00 and includes more complex devices. In article <22093.9502011123@bmth.ac.uk> mvasilko@bournemouth.ac.uk (Milan Vasilko) writes: >Hello ! > > >I wonder if anyone has any information about: > >- commercially available "on-fly" reprogrammable devices, i.e. devices > which can be programmed during their operation (or better part of the > chip can be reprogrammed while another parts are still operational).Article: 674
I am currently working on a board that will use XC4013's and possibly a XC4005H and/or XC4003H. I was thinking about using the Xchecker cable supplied by Xilinx to configure the chips on the board until the design is stable enough to warrant using PROM's. Has anyone ever tried this? Is it possible/worhtwhile? Specifically, is it possible to configure multiple chips in daisy-chain fashion using the Xchecker cable? Would using the JTAG port be a better option? We are also planning on using the boundary scan capability to check the design. However, we are having difficulty finding suppliers of the equipment necessary, especially since we are still not sure exactly what equipment we would need ( Hey, we're only students after all ). Does anyone know what companies I could contact for catalogs/databooks/information? Thanks in advance for any advice. David Hoffmeister University of Buffalo
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