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
Also http://www.nallatech.com/ Ballynuey3: High Performance PCI DIME Motherboard With Xilinx VirtexT-II http://www.nallatech.com/products/dime_professional/ballynuey3/index_en.htmArticle: 33026
Hi Alex, Solution for problem(1) and (2): Inserting the tristate Buffers BUFT(x) to all the signals, which are connected to the BUS, then recompile it, you might see the warning messages like "WARNING : (HDL__0001). Signal <XXXX(y)>> has a multisource." Just ignore it, that's all. BTW, which Xilinx Foundation are you using? V3.3i? Hope this helps, HongArticle: 33027
as far as i know, the the spartan-2 development board is one of the best and cheapest i have seen (the one from burch electronic designs). it sounds to me like you are worrying about too many details, when instead, you should be writing code and making sure it will work inside any fpga. if you are truly just starting, you should concentrate on writing synthesizable code first, since this concept is the hardest to understand for newbies. then, you should be able to create a test platform to exercies your code, ie: testbench. if you cannot get this far, then having the latest and greatest fpga, or even cpld, etc... will be of no use to you. if you do write the code and synthesize it, then there is still many things you need to do such as place and route it, analyze the timing, constrain the design, and finally the last step is the download. if you want to do a design, know what you are designing. if you just want to do a networking project, then be more specific than network switching, since network switching is actually quite complex. anyways, the main point is the downloading the fpga is the last thing you want to do in your design. the bulk of the time is spent making it work in your simulation environment since finding a bug in simulation is an order of magnitude easier than debugging in the hardware environment. just do the design and then you will know enough about fpgas to know what you want. chrisArticle: 33028
In article <3B51E407.80E2FE21@mail.com>, johnhandwork@mail.com says... > Say, can we still buy TTL? DTL? RTL? Sure. Iv'e got some stuff with '72-'74 date codes on 'em... Wanna buy? ;-) ...other than that, it's the CMOS stuff and much of that is hard to come by. ---- KeithArticle: 33029
Suhaib, > I've seen your plug many times before! While it seems quite good, I Thanks :) > don't like the fact that it's all on ribbon cables. Specifically, I > worry that the download cable bit will break. Does it include JTAG > connections? > The cables are all really robust. I have some units here that have been used alot, will all kinds of pulling, twisting, plugging, unplugging etc etc. Not a single breakage or failure in all that time. The main B3-SPARTAN2+ board has two download header connectors on it. One for JTAG download, and one for serial mode download. So you can choose your preferred download method, JTAG or serial mode. The three download mode pins M0, M1 and M1 are jumper programmable on the board. Best regards Tony Burch http://www.BurchED.com.au Lowest cost, easy-to-use FPGA prototyping kits! > To everyone else: > > What is the difference between Spartan-II and Virtex-II, what > different markets are they targeted at, and what are the different > features? Why chose one over another? > > Thanks. > > > > "Tony Burch" <tony@BurchED.com.au> wrote in message news:<3b50f1d9@news1.idx.com.au>... > > Suhaib, > > > > You may wish to consider the low cost > > SpartanII board from Burch Electronic Designs: > > http://www.burched.com.au/bedspartan2.html > > > > It has a 200K gate SpartanII device, configuration > > download pod and cable, a header programmable > > PLL oscillator (1-100MHz), and it works with > > the free Xilinx WebPACK design software. > > All for less than US$120 ! > > > > Best regards > > Tony Burch > > http://www.BurchED.com.au > > Lowest cost, easy-to-use > > FPGA prototyping kits!Article: 33030
Hi, anilkateel@rediffmail.com (Anil) wrote in message news:<f1a78ec4.0107132313.518c8b1@posting.google.com>... > There is no search path concept in FPGA Express/Compiler. > > So when you include a verilog file using `include directive > you need to make sure that the file is in the same directoy > where as the original verilog file exists. > > There are 3 solutions for this :- > (1)If suppose u have a file /a/b/c/test.v which has a `include x.v > and if x.v is not there under /a/b/c/ DIRECTORY then just create > a link to x.v under /a/b/c/ pointing it to it's original location. > link -s <original x.v path>/x.v /a/b/c/x.v > (2)Copy x.v to /a/b/c/ directory. > (3)Just chnage `include x.v to `include <locattoin to x.v>/x.v (Hard to do !!!) > Is +incdir option supported by FPGAExpress? It is a defacto standard option that directs the Verilog compilers where to look for the included files. SriniArticle: 33031
Good Morning, my question is this : Matlab design for me a polyphase filter using the function rcosflt , the number of coefficient produced is function of the interpolation rate (...why ??) in particular for interpolation 3 I've 19 coefficients, for interpolation 4 I've 25 coefficients and for interpolation 6 I've 37 coefficients . This means that I always have 1 fir with 7 taps and (interpolation_rate -1) filters with 6 taps , so I had think to produce only one polyphase filter that interpolate 6 and when I want to interpolate 3 I use the first 3 FIR, when I want to interpolate 4 I use the first 4 FIR and when I want interpolate & I use all 6 FIR , naturally the coefficients are programmable. I would want to ask you if this is correct by your point of view ... Thank you in any case ... Antonio D'OttavioArticle: 33032
The DDS core is not in the evaluation CD but it is on the site that tell me that I could have it with the evaluation CD, Do you know well how can I have this core ?? Thanks in advance .. Antonio D'OttavioArticle: 33034
In article <ee7173c.0@WebX.sUN8CHnE>, Hong <lvhong@aol.com> wrote: >Hi Alex, >Solution for problem(1) and (2): >Inserting the tristate Buffers BUFT(x) to all the signals, which are connected > to the BUS, then recompile it, you might see the warning messages like > "WARNING : (HDL__0001). Signal <XXXX(y)>> has a multisource." Just ignore it, > that's all. Well, I would have thought that I should be able to ignore it, but the implementation refuses to compile, even though I have as you suggest tristate buffers at the outputs (how else, I wonder, would you *want* to implement a bus?) And it doensn't appear, through an examination of the log, that anything else is wrong. So it does at least seem as though Xilinx doesn't like it, regardless of what it should like. > >BTW, which Xilinx Foundation are you using? V3.3i? 2.1i at the moment. We bugged our rep a while back to get us the 3.3i software, but so far it hasn't arrived. I have been on them a few times about this. It didn't help that the main FAE at our XIlinx dist'y moved on to another job. The new guy is going to take a while getting up to speed on what his customers need. Alex Rast arast@qwest.net arast@inficom.comArticle: 33035
Hi everyone, I am using foundation 3.1 with service pack 8, and IP core4. I can't create a distributed Memory4.1 (Memories & Storage Elements/RAMs & ROMs) for a Virtex2. ERROR: Unable to open file for Memory Initialization: D:\Projets\memory.mif ERROR: An internal error has occurred. To resolve this error, please consult the Answers Database at http://support.xilinx.com ERROR: Sim has a problem implementing the selected core. Implementation netlist will not be generated. ERROR: SimGenerator: Failure of Sim to implement customization parameters core memory ERROR: Core memory did not generate EDIF implementation netlist (.EDN) file. ERROR: The .XSF symbol generation support file cannot be generated because it requires an implementation netlist. WARNING: Warnings and/or errors encountered while generating memory (Distributed_Memory 4.1) All output products requested may not have been generated. ERROR: Elaboration failure for core Distributed_Memory ERROR: Elaboration of core Distributed_Memory failed. Does any body knowing way how to avoid that problem? Thanks Benoit.Article: 33036
for information ... http://www.freecore.com/nosupport/blaster.htmArticle: 33037
Ok, I am looking for a good introductory book, not to VHDL, but more to the whole shebang behind design of FPGAs. I've done courses in college on Digital Electronics, and some practicals with Altera PLDs, but I'd like a way of putting things together. Questions like: -How to design with the target architecture in mind -Good and bad practice -Efficient use of HDL -Interfacing with other devices -Timing considerations The more modern the book, the better. I'll probably be getting Ashenden's "Designer's Guide to VHDL"/2e anyway, so I need something which would be a good complement. If it covers Xilinx chips, even better. Also, what resources might I use to get in tune with my target Architecture? Is it just a matter of data sheets and experience? The sorts of terms puzzling me: -Block RAM - how do you use these? -Distributed RAM -DLLs - what are they for? -Cascade Chains -Different types of clocks (I know wat a clock does, but what're all the different types?) As you can see these are the questions of an ignorant beginner. Hence the sort of book I need. Thanks all.Article: 33038
"Keith R. Williams" wrote: > In article <3B51E407.80E2FE21@mail.com>, johnhandwork@mail.com > says... > > Say, can we still buy TTL? DTL? RTL? > > Sure. Iv'e got some stuff with '72-'74 date codes on 'em... > Wanna buy? ;-) > Maybe you should consider embedding them in perspex (plexiglass) and selling them as charms to new age h/w designers to ward off the evil metastability demon. I'd buy one to ward off s/w engineers who think a critical path is the route to the nearest pub (bar).Article: 33039
Sould be no problem if the dsp has no problems with tri state pins. But how are the config pins driven on the board original? You could get a contention with the byte blaster. Martin -- Whant to see the evolution of a Java processor? http://www.jopdesign.com "Russell Shaw" <rjshaw@iprimus.com.au> schrieb im Newsbeitrag news:3B4D8E87.31046A07@iprimus.com.au... > Hi all, > > I've got a dsp driving an altera ep1k30, and was wondering if its > ok to reprogram that chip even tho the dsp is still driving it. > I've got a byteblaster connected to it. I read somewhere that > all(?) the pins go to tri-state when programming (passive serial). > > -- > ___ ___ > / /\ / /\ > / /__\ / /\/\ > /__/ / Russell Shaw, B.Eng, M.Eng(Research) /__/\/\/ > \ \ / Victoria, Australia, Down-Under \ \/\/ > \__\/ \__\/ >Article: 33040
Hi all, To query if some peice of hardware is present :- I need to write a known bit pattern of length n=8 to a card and read the loop back from and if i get the same pattern back....that indicates the card is present. if its not the same indicates the card is not there... So, how should i go about this..... Many thx for ur help.... -- Abhimanyu Rastogi 2nd year student University of Ottawa CanadaArticle: 33041
That one's old hat with the 'LS244 and all... PYD wrote: > > for information ... > > http://www.freecore.com/nosupport/blaster.htmArticle: 33042
Subodh Nijsure wrote: > > > Now the DONE pin goes high, and I my FPGA is seems to be doing its job. > But what I have observed is, as soon as checksum command words are sent > for the last block /INIT pin goes low, but if ignore that error, DONE > pin is still going high and FPGA seems to be working okay. > > I haven't tried pulling /INIT high as was suggeseted earlier. Any hints > on why /INIT goes low? > > /Subodh Nijsure just an idea: INIT has a weak pullup before and while configuration (?). When configuration is done, INIT becomes a user I/O-pin, and the pullup is disabled (?). As you don't use INIT as user I/O, the pin is at high impedance state. Your scope probe forces it to low. WernerArticle: 33043
This question seems to fall somewhere between circuit design and compiler design, so I'm hoping the denizens of this group may be able to address it. This expression is "clearly" replaceable by a single "false" (a eq 10 and a eq 30) Is there a name for this optimisation? Is it commonly implemented? BugBearArticle: 33044
pwomack@engage.com (bugbear) writes: > (a eq 10 and a eq 30) > > Is there a name for this optimisation? Constant folding ?Article: 33045
In article <3B5053B0.8E3FF4D2@ecubics.com>, emu@ecubics.com says... > Noddy wrote: > > > > This brings me to my question? Should I rather be > > attempting to implement my designs in VHDL instead? > > Mix it ! > A top level schematic, the units below in VHDL, Verilog, etc.. > It's amazing, already magic how easy it is to explain what you're doing > to somebody else, when you can see all functional units, data flow on > one sheet. > And, sometimes you see your own design more clearly ;-) That certainly would be an advantage. However, I understand that synthesis in a mixed environment doesn't work well. The redundancy isn't squished out beyond the black boxes. > when you like to go to HDL only, just replace one sheet to a text page. If I could do both schematics and (V)HDL with the same toolset with the flattening occurring before the "real" synthesis this would be nice. This soul also allow us old fuddy-duddy schematic wonks pick up on HDLs wile remaining productive. Listening Synplicity? ;-) ---- KeithArticle: 33046
Ray Andraka schrieb: > > but including Spartan II doesn't make sense. For new designs, if you are I think it makes sense, since the Spartan-II devices are cheaper. > not in the Spartan II, I would go with the Virtex E right now. The Virtex As I said. > design into space, you're stuck with Virtex. Personally, I'd forgo the > 4K/Spartan (original) architecture for new designs, as the cost per gate and > speeds are better for the later families, and the later families have a Yes, but there are lots of applications that dont need the Multi-million gate FPGAs. If you have some small on-on boards which needs just some glue logic, small fifos, some uC registers etc., Spartan(XL) is just fine and cheap. Not everyone is working everyday on cutting edge, breakthrough designs. ;-) > longer time horizon until they become obsolete. Yes. -- MfG FalkArticle: 33047
sounds like you should probably go to xilinx support pages and check out the library. download all the docs that you have questions about, print them out, and go to your local coffee shop to study them. no book i know of has the recent developments of the ins and outs of fpgas, especially the newer xilinx ones. if you really must have a book, probably get "Real World FPGA Design with Verilog", assuming you don't mind learning verilog. that book is quite useful for the basics, but it seems like fpgas have gone a long way since then. chrisArticle: 33048
Hi, Is there a simple way to fix manual or automatic routing so that on the next PR cycle it would not change ? The only way I know to do this is by instantiating predefined macro and I am looking for an easier way. Thanks, Rotem. Rotem Gazit MystiCom LTD mailto:rotemg@mysticom.com http://www.mysticom.com/Article: 33049
I have a .bit file that I converted to a .hex file with PromGEN. When I look at this .hex file, I noticed there is only one 32 bit dummy word before the syncronization word. I read in XAPP176 (from the Xilinx website) that there are supposed to be 2 32 bit dummy words. Has anyone else had similar problems? Just curious, because we can't get this .bit file to load into the FPGA. Is there a definitive description somewhere of the header that a .bit file should have for a Spartan II? Thanks, Dave
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