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
Tim Hubberstey <tim_hubberstey@mindlink.bc.ca> wrote in article <tim_hubberstey.95.32F030CF@mindlink.bc.ca>... > I recommend that you look at the Synopsys FPGA compiler for the PC (I can't > remember the exact name). We do our design work in VHDL and until recently had > been using Viewlogic but we found it to be such a pain to work with (both > Windows and Unix versions!) that we dumped it completely and switched (at > great expense) to Synopsys. Our gate counts immediately dropped by at least > 30% and the timing improved by a similar margin. > The PC equivalent of the Synopsys FPGA compiler is called FPGA Express. You can find out more information at: http://www.synopsys.com/products/fpga_pc/fpga_pc.html Synopsys is quite liberal with demo CD-ROMs and it's definitely worth a look. For us non-university folks, it about $13,000 list price with one language supported and one device optimizer. -- Steven Knapp E-mail: optmagic@ix.netcom.com Programmable Logic Jump Station: http://www.netcom.com/~optmagicArticle: 5201
I am considering the design of a dynamically reconfigurable logic platform. It is more in line with tinkering than a serious product application. Has anyone else tried to build an in circuit programmable interface inside an FPGA or CPLD? I realize that Lattice offers something of this sort(pseudo JTAG port) but I want to go one step further; to make connections between "user defined" muxes surrounding "user defined" logic blocks. In short I want to waste the resources provided for me by the chip manufacturer and third party tool designers to functionally design and program an FPGA or CPLD in favor of using those resources to build a "scaffolding" inside to facilitate the construction of dynamically reconfigurable logic. I realize this is wasteful. I want control over routing delays and repeatability without consulting a third party. It also would allow greater flexibility in interfacing to C compiler outputs. Just curious. . .Article: 5202
It is not clear to me what this posting really means, but I might point out that all Xilinx FPGAs are configured by loading bits into latches, and they can therefore be re-configured on the board within milliseconds and an unlimited number of times. Also, the XACT development system ( you need it, at least the simplest and cheapest version of it) allows you to tinker with the available resources to your heart's content. Any feature that the silicon is able to do is also available to you. For detailed product information, look at our web site ( www.xilinx.com ) Peter Alfke, Xilinx ApplicationsArticle: 5203
I've been using the Altera FPGAs for 6 months or so and one of the first things that bugged me was the cost of the BitBlaster. $200 for a little box (with an Altera FPGA in it no less) just to download the parts. So I dabble in programming and wrote a short program to wiggle a couple of bits on the printer port to download my 81188 and 10K50 parts. It works great and is at least as fast as the BitBlaster. It reads the design's .ttf file created when you do place and route. Any interest? I'll e-mail or post depending on how many responses I get.Article: 5204
>And bye-bye to support if you need it, since most companies don't want to >support old versions of software when bugs have been fixed in later versions. This applies to all software eventually. The stuff you buy today will be unsupported in say 4 years' time, just like the stuff you bought 4 years ago is unsupported now. Nothing ever changes. Is comes down to a financial and time decision: do you want to keep paying the annual "maintenance" to always have the latest, and do you then want to put in the time to learn of the changes, discover new bugs - even when the work you do may not need the latest version. But this is an old argument... Peter. Return address is invalid to help stop junk mail. E-mail replies to z80@digiserve.com.Article: 5205
Rick Filipkiewicz wrote: > Also: Do any of you Xilinx users out there know where I can get a > description of the XNF file syntax ? Or is this still a deep dark > secret? The xnf spec version 6.1 is on our FTP site at : ftp://ftp.xilinx.com/pub/documentation/xnfspec.pdf And the 7K and 9K family addendum which sound like what you are interested in : ftp://ftp.xilinx.com/pub/documentation/xnf61_7k.pdf Finally, postscript versions of these files are at : ftp://ftp.xilinx.com/pub/documentation/xnfspecs_61.tar.Z -- ---------------------------------------------------------------------------- / 7\'7 Brian Philofsky (brian.philofsky@xilinx.com) \ \ ` Xilinx Technical Applications Engineer hotline@xilinx.com / / 2100 Logic Drive 1-800-255-7778 (toll-free) \_\/.\ San Jose, California 95124-3450 1-408-879-5199 ----------------------------------------------------------------------------Article: 5206
Ed Vogel wrote: > > I am considering the design of a dynamically reconfigurable logic > platform. It is more in line with tinkering than a serious product > application. Has anyone else tried to build an in circuit programmable > interface inside an FPGA or CPLD? > > I realize that Lattice offers something of this sort(pseudo JTAG port) > but I want to go one step further; to make connections between "user > defined" muxes surrounding "user defined" logic blocks. In short I want > to waste the resources provided for me by the chip manufacturer and third > party tool designers to functionally design and program an FPGA or CPLD > in favor of using those resources to build a "scaffolding" inside to > facilitate the construction of dynamically reconfigurable logic. I > realize this is wasteful. I want control over routing delays and > repeatability without consulting a third party. It also would allow > greater flexibility in interfacing to C compiler outputs. Just curious. . > . Well....:) I just read a simular paper where the design was reconfigurable switches and 2-bit processors. You want to prototype a FPGA in an FPGA. Sounds very feasible for a small device. Now if you wanted to emulate an emulator that might be a different story. You might look to my newest most favorite page http://dec.bournemouth.ac.uk/dec_ind/decind6/drhw_page.html For all you old timers: check this page out it is really very good. Steve Casselman http://www.vcc.com/hotann.html The IBM PC of Reconfigurable Computing (TM)Article: 5207
Dear everybody, Can any expect here tell me what is the different between FPGA and CPLD? Thanks -- Thank you for your kind attention. SAM CHENGArticle: 5208
Whould the person who last posted the address of Michael J. Wieners paper "Efficient DES Key Search" please post the address of that paper. I have a printout of hte paper but no address to let others know where to pick it up. The Challenge is on and the 40-bit key has been broken ($1,000) The next key is 48-bits. The paper has a gate level schematic in it and says it takes 2500 registers for a fully pipelined version. Steve CasselmanArticle: 5209
I think you should also look at Exemplar's Galileo. it's VITAL compliant and is integrated with timing analysis and MTI's V-System simulator. It supports all the major programmalbe vendors. University deals are available. See their web page at www.exemplar.comArticle: 5210
Peter (zx80@dgiserve.com) wrote: : > Orcad with Xilinx library and Xact : Interesting. Was this SDT/386? And how did you manage with simulation? I did not. : The main reason I am keeping the ancient Viewlogic stuff is that it : has a good simulator. My experience of VST was very bad. I knew it and did not even try. I switched to ViewLogic about 4 years ago. George.Article: 5211
Wayne Turner (waynet@goodnet.com) wrote: : WRONG!!!!! There is NO receptionist at the Altera help-line, at least in the : United States. You will be on hold until someone is available Excellent!! Just 5 or 10 or 20 minutes on hold and that's it. : and the person who comes on the line is an applications engineer, NOT a : receptionist. And then he plays the usual routine : "Please-send-me-your-code-and- I-will-take-a-look-at-it". Why waste time? I can do it before he asks. In my experience email is more effective than phone calls provided the customer support uses it properly. Last time I checked (about a year ago) Xilinx did it better. BTW another company that gave me very good support by email was Atmel. They have much less resources than Altera or Xilinx but they did their best. : What is their option? Build a fab to make FlashLogic? They are put in the : same situation as you are. Obsolescence happens. They could absorb part of the costs. Or think twice before they bought the part. Or think three times before they converted it from EPROM to FLASH. : > Yes, but this part never got out of "preliminary" category and there are : > no customers that were using it for years. : Does it matter how long people were using it? Yes. The longer the part is available the more designs are utilizing it. : If it does, how long had people : been using FLASHlogic, considering it has only been around for a few years : (and owned by Altera for less than two)? For our company it was long enough to use it in about half of our products. : No, it wasn't. If your wafer supplier won't make the wafer anymore, what can : you do if they are the only fab with that technology? Don't make customers pay for it. Especially if there are so few of them as we are being told. George.Article: 5212
man cheng wrote: > > Dear everybody, > > Can any expect here tell me what is the different between FPGA and CPLD? > Thanks > -- > Thank you for your kind attention. > > SAM CHENG Simplistically speaking, the difference is in the architecture. An FPGA is an array of relatively simple configurable logic cells. The smallest arrays are about 8 x 8 cell matrices. Each cell in the FPGA typically has a small number of inputs (2 to 9 depending on the device) and somewhat limited routing between cells. Each cell has typically has at least one flip flop (there is an exception that uses a combination of two adjacent cells to construct a flip-flop). The resulting architecture is rich in registers, but is not well suited for wide high speed combinatorial stuff. In contrast, the CPLD structure is generally more like that of a traditional PAL, where each "macrocell" consists of a register and some form of an and-or logic array. The number of macro-cells is small compared to the number of logic cells in an FPGA, but the cells can handle fairly wide combinatorial functions with no speed penalty. The macrocells are usually interconnected with some form of global routing resource. Hope the short answer is helpful. -Ray Andraka, P.E. Chairman, the Andraka Consulting Group 401/884-7930 Fax 401/884-7950 email randraka@ids.net http://www.ids.net/~randrakaArticle: 5213
Hi all, Can anybody give me a hint of what's wrong with my newsreader? Every time I try to open a message from this guy, my newsreader (netscape3.0) says: No Such article, Perhaps the article has expired! He is the only one treated this way! Best Regards Aage Farstad aage.farstad@ffi.noArticle: 5214
Ray Andraka (randraka@ids.net) wrote: : man cheng wrote: : > : > Dear everybody, : > : > Can any expect here tell me what is the different between FPGA and CPLD? : > Thanks : > -- : > Thank you for your kind attention. : > : > SAM CHENG : Simplistically speaking, the difference is in the architecture. An FPGA : is an array of relatively simple configurable logic cells. The smallest : arrays are about 8 x 8 cell matrices. Each cell in the FPGA typically : has a small number of inputs (2 to 9 depending on the device) and : somewhat limited routing between cells. Each cell has typically has at : least one flip flop (there is an exception that uses a combination of : two adjacent cells to construct a flip-flop). The resulting : architecture is rich in registers, but is not well suited for wide high : speed combinatorial stuff. In contrast, the CPLD structure is generally : more like that of a traditional PAL, where each "macrocell" consists of : a register and some form of an and-or logic array. The number of : macro-cells is small compared to the number of logic cells in an FPGA, : but the cells can handle fairly wide combinatorial functions with no : speed penalty. The macrocells are usually interconnected with some form : of global routing resource. : Hope the short answer is helpful. : -Ray Andraka, P.E. : Chairman, the Andraka Consulting Group : 401/884-7930 Fax 401/884-7950 : email randraka@ids.net : http://www.ids.net/~randraka Probably the major difference between the two types is that CPLDs, or ``fat PALs'', have very simple and predicatable pin-to-pin timing. This means that, within fairly tight bounds, recompiling & rerouting a design won't cause speed varitions in the resulting device. FPGAs on the other hand have a large amount of internal resource and variable routing delays. Keeping a given performance level over design changes can be difficult and is highly dependent on the efficacy of the vendor's floor planning and place/route tools. I think a rule of thumb might be to use FPGAs for ``rigid'' devices like data paths, arithmetic units etc. or for non-speed critical bits of random logic. CPLDs are better suited to the high clock rate control logic that might have to be rebuilt frequently during the prototype phase. CPLDs are also a lot cheaper! _________________________________________________________________________ Dr. Richard Filipkiewicz phone: +44 171 700 3301 Algorithmics Ltd. fax: +44 171 700 3400 3 Drayton Park email: rick@algor.co.uk London N5 1NU EnglandArticle: 5215
I'm planning to jump into the programmable logic design arena in order to implement DSP and FEC algorithms that surpass the capability of programmable DSPs. I'm going to follow the Altera family of devices and have ordered their development tools (MAX-II Plus): graphic entry, VHDL input option, compiler, simulator, timing analyzer etc. I've seen from this newsgroup that a large number of people use a tool called Viewlogic. What is the difference between the tools offered by Altera and those from Viewlogic. Thanks. Richard Young Signal Processing Engineer Communications Research Centre, Ottawa, Canada richard.young@crc.doc.caArticle: 5216
Hi Steve, I'm not sure it is the easiest way of getting a copy of this paper, but I just downloaded the 107 kBytes gzipped postscript file from Japan at the following address: http://www.aist-nara.ac.jp/Security/doc/paper/wiener.deskeysearch.ps.gz Sure this will help! Salut Chris. -- Christophe Beaumont ( http://ubolib.univ-brest.fr/~beaumont ) Laboratoire d'Informatique de Brest Projet ArMen ( http://ubolib.univ-brest.fr/ )Article: 5217
Job Title: ASIC DESIGN ENGINEER Job/Skill Requirements: BSEE or MSEE. Must have 2+ years experience with ASIC design, Should have experience with Synopsis and Verilog or VHDL. FPGA experience is a plus, Knowledge of C/C++, MPEG2, or ATM is a plus. Job Description: Development, design, modification, & verification of complex digital integrated circuits and FPGA’s for video & data services. Product architecture, development of design methodology, circuit design & prototype debugging. Compensation: 45k to 70k, depending on experience Duration: Permanent Jobsite location: Atlanta, GA Start date: ASAP For the above position, please respond by telephone & fax resume to: NOTE: To ensure receipt of your resume, it must be ASCII format __ Anthony Dozier _________________ (Placement Specialist) Internet: anthony_dozier@systemone.com Fax: 404-252-0073 Phone: 404 255-5004 x105 System One Technical 5775 Peachtree Dunwoody Rd Suite B220 Atlanta, GA 30342Article: 5218
In article <32F1173E.43A8@tisc.com>, Steve Schossow <ss@tisc.com> wrote: >I've been using the Altera FPGAs for 6 months or so and one of the first >things that bugged me was the cost of the BitBlaster. $200 for a little >box (with an Altera FPGA in it no less) just to download the parts. > >So I dabble in programming and wrote a short program to wiggle a couple >of bits on the printer port to download my 81188 and 10K50 parts. > >It works great and is at least as fast as the BitBlaster. It reads the >design's .ttf file created when you do place and route. > >Any interest? I'll e-mail or post depending on how many responses I >get. Altera now has a ByteBlaster that does the same thing. It is (and yours should be as well) quite a bit faster than the BitBlaster since the BitBlaster was limited by the UART in your PC/workstation, which is quite a bit slower than a parallel port. They still charge you $150 for the ByteBlaster though, just so you know where yours might fit in. WayneArticle: 5219
In article <32f7f5b2.24840008@news.alt.net>, zx80@dgiserve.com (Peter) wrote: > >>And bye-bye to support if you need it, since most companies don't want to >>support old versions of software when bugs have been fixed in later versions. > >This applies to all software eventually. The stuff you buy today will >be unsupported in say 4 years' time, just like the stuff you bought 4 >years ago is unsupported now. Nothing ever changes. > >Is comes down to a financial and time decision: do you want to keep >paying the annual "maintenance" to always have the latest, and do you >then want to put in the time to learn of the changes, discover new >bugs - even when the work you do may not need the latest version. > >But this is an old argument... True. Both what you said at the beginning AND the fact that it is an old argument ;) WayneArticle: 5220
In article <5cs4oq$cf5@borg.svpal.org>, garyk@svpal.svpal.org (George Noten) wrote: >Wayne Turner (waynet@goodnet.com) wrote: > >: WRONG!!!!! There is NO receptionist at the Altera help-line, at least in the > >: United States. You will be on hold until someone is available > > Excellent!! Just 5 or 10 or 20 minutes on hold and that's it. My experience has been less than five minutes. And by the way, what happened to your "receptionist" claim? >: and the person who comes on the line is an applications engineer, NOT a >: receptionist. > > And then he plays the usual routine : "Please-send-me-your-code-and- > I-will-take-a-look-at-it". Why waste time? I can do it before he asks. > In my experience email is more effective than phone calls provided the > customer support uses it properly. Last time I checked (about a year > ago) Xilinx did it better. BTW another company that gave me very good > support by email was Atmel. They have much less resources than Altera > or Xilinx but they did their best. The point of the help desk is not tell you how to write code; go to college for that. The point is to help people with specific issues in using the tools or devices. If you have code that is crashing the software, then they want it so they can fix it for next release. Most of the issues I had to call about were resolved over the phone at the time that I called. >: What is their option? Build a fab to make FlashLogic? They are put in the >: same situation as you are. Obsolescence happens. > > They could absorb part of the costs. Or think twice before they bought > the part. Or think three times before they converted it from EPROM to > FLASH. 1. They are absorbing part of the costs by making a large end-of-life buy on wafers. 2. Do you think they would have bought the part if they knew 2 years down the road that Sharp wasn't going to want to make it anymore? 3. Converted? What the hell are you talking about? It was FLASH when Intel owned it, it was just called FlexLogic. When Altera bought it they had to change the name because they already had the FLEX 8000 family and didn't want it to be confused. So they named the part according to what it really was: FLASH. Hence the name FlashLogic. Perhaps you should come back after you've read up a bit. >: > Yes, but this part never got out of "preliminary" category and there are >: > no customers that were using it for years. > >: Does it matter how long people were using it? > > Yes. The longer the part is available the more designs are utilizing it. And how long was it out? >: If it does, how long had people >: been using FLASHlogic, considering it has only been around for a few years >: (and owned by Altera for less than two)? > > For our company it was long enough to use it in about half of our products. What is the product life? The parts will be available to be shipped for another year and a half. Is that not long enough notice? >: No, it wasn't. If your wafer supplier won't make the wafer anymore, what can >: you do if they are the only fab with that technology? > > Don't make customers pay for it. Especially if there are so few of them as > we are being told. You didn't answer the question. WayneArticle: 5221
Hello all Altera designers! I would like to promote the idea of a FreeCore Library. The FreeCore Library consists of free, parameterizable building blocks for Altera programmable logic that can be used in your design for absolutely free! My first contribution is the parameterizable Compact UART. This is a simple, complete and compact UART. An 8-bit UART may consume as little as 63 logic cells in a FLEX 8000 or FLEX 10K architecture, including transmitter, receiver and baud rate generator. The documentation and download page is available at: http://www.geocities.com/SiliconValley/Lakes/3656/ Please give me your feedback on the FreeCore philosophy. I plan to set up a FreeCore site soon, and I hope that you, the designers and application guys, would like to contribute to the library. What kind of functions would you like to see in the FreeCore library? I'm planning my next release soon: A single master I2C (I squared C) controller. NOTE: Although I work as a Field Application Engineer for one of the Altera distributors, I would like to state that this is totally my personal initiative. Altera is not involved in any way.Article: 5222
Hello all Altera designers! I would like to promote the idea of a FreeCore Library. The FreeCore Library consists of free, parameterizable building blocks for Altera programmable logic that can be used in your design for absolutely free! My first contribution is the parameterizable Compact UART. This is a simple, complete and compact UART. An 8-bit UART may consume as little as 63 logic cells in a FLEX 8000 or FLEX 10K architecture, including transmitter, receiver and baud rate generator. The documentation and download page is available at: http://www.geocities.com/SiliconValley/Lakes/3656/ Please give me your feedback on the FreeCore philosophy. I plan to set up a FreeCore site soon, and I hope that you, the designers and application guys, would like to contribute to the library. What kind of functions would you like to see in the FreeCore library? I'm planning my next release soon: A single master I2C (I squared C) controller. NOTE: Although I work as a Field Application Engineer for one of the Altera distributors, I would like to state that this is totally my personal initiative. Altera is not involved in any way.Article: 5223
=?iso-8859-1?Q?Bj=F8rn?= B. Larsen wrote: > > Aage Farstad wrote: > > > > Can anybody give me a hint of what's wrong with my newsreader? Every > > time I try to open a message from this guy, my newsreader (netscape3.0) > > says: No Such article, Perhaps the article has expired! He is the only > > one treated this way! > > The same happens to me! > > I am also using Netscape 3.0 now. > > Bjørn BL. > I noticed it too, with netscape 2.?. I did manage to see one of his articles using trn on a UNIX box, and it has a special character in the message-id field. (No, I don't recall exactly what: perhaps one of {#$%}). I could tell that was the problem, because the netscape error message indicated the message-id, and the indicated message-id was truncated starting with the offending character. So, Steve, if you want Netscape users to read your messages, see if you can somehow get your software and/or ISP to provide a message-id. I suspect that somewhere there is an Internet RFC specifying the grammar for parsing this field, that most newsreader-writers decided to be generous in accepting errors, and that Netscape is being technically correct (but wrong from a software engineering viewpoint) in rejecting them. Perhaps this should be further considered somewhere in the news.software.*, or news.admin.* hierarchies. -- R. T. Wurth / (w) Holmdel, NJ / (h) Rumson, NJ rwurth@att.comArticle: 5224
Any suggestions to a problem I have would be appreciated, > > I support a LAN with an FDDI backbone, connected by 4 ESXMINs around the campus. We have a cisco 4500 series router which has an FDDI port but is not directly connected to the backbone at the moment. The router routes IP, and bridges IPX and Appletalk. The routing protocols we are using areRIP and OSPF but would like to solely use OSPF eventually. My problem is > I want to connect the router directly into the FDDI backbone, but when I do it will not bridge the appletalk over the backbone, although there are no problems with the IP and IPX. -- Travis Terry Network Support Officer Charles Sturt University Ph:(069) 332817 Fax:(069)332454 email:tterry@csu.edu.au homepage:http://golum.riv.csu.edu.au/~tterry/ Do we seek advice to confirm what we have already decided?.......
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