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
DSP is Digital Signal Processing or Processor. Usually people automatically jump to thinking a DSP microprocessor, which is a microprocessor whose instruction set and bus architecture are optimized somewhat for signal processing applications. These generally have a single cycle multiply as well as efficient looping constructs to streamline inner loops and often special addressing modes. Digital signal processing really encompasses much more hardware than the traditional DSP microprocessor. By using special purpose (custom) hardware, you can get much higher data rates than a microprocessor is capable of. That's what I do with FPGAs. DSP microprocessors are every where....modems, graphics cards, sound cards, cell phones, CD players, disk drives and so on. "Jean-Réginald Louis" wrote: > Excuse me! I didn't have a clear idea of what it's a DSP. In my course, I'm > learning microcontroller. I have friends studying telecom (college level) > and they using DSP to manipulate signal (like sound) and I automatically > made an association between telecom and DSP. > > Can someone tell me the difference between a MCU (microcontroller), a DSP > and a CPU. What type of processing unit are use in what type of > project/environement/situation/etc ? Why a graphics card use a DSP (like the > nVidia GeForce (if I'm not wrong)) ? > > > Hi! > > > > DSP is *Digital* signal processor or processing. > > > > You can not handle analog data in with an FPGA, you first have to > > convert the analog signal to a digital signal and then process it. > > > > / Jonas > > > > On Mon, 13 Mar 2000 00:15:14 -0500, "Jean-Réginald Louis" > > <louis.reginaldjean@teccart.qc.ca> wrote: > > > > >Hi. I'm new to this field. I have a good idea of what we can do with FPGA > > >but I saw something that I can understand. I found a companie (in the > Web) > > >that do DSP with FPGA. If I'm not wrong DSP are like MCU, but > specifically > > >designed to handle analog data. So how can we handle analog data in FPGA > > >pins. What's worst, I saw (in the same compagny) "Modulation and > > >demodulation with FPGA". How can we do this? > > > > > > > > > > > > > > Hi! > > > > DSP is *Digital* signal processor or processing. > > > > You can not handle analog data in with an FPGA, you first have to > > convert the analog signal to a digital signal and then process it. > > > > / Jonas > > > > On Mon, 13 Mar 2000 00:15:14 -0500, "Jean-Réginald Louis" > > <louis.reginaldjean@teccart.qc.ca> wrote: > > > > >Hi. I'm new to this field. I have a good idea of what we can do with FPGA > > >but I saw something that I can understand. I found a companie (in the > Web) > > >that do DSP with FPGA. If I'm not wrong DSP are like MCU, but > specifically > > >designed to handle analog data. So how can we handle analog data in FPGA > > >pins. What's worst, I saw (in the same compagny) "Modulation and > > >demodulation with FPGA". How can we do this? > > > > > > > > > > > -- -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: 21251
Kindly accept an invitation to participate in a survey about your interest in and use of books on computer and electrical engineering. Participating will involve spending about twenty to thirty minutes taking an online survey. You'll need to answer a few qualifying questions first. If you qualify and complete the survey, we will reward you with your choice from a selection of books on computer and electrical engineering. If you are interested, please click here http://survey.informative.com/sb/survey?s=62000&hdr=1&xu=NGCAF Thank you for your time and interest. Sent via Deja.com http://www.deja.com/ Before you buy.Article: 21252
Hi. Would anyone know how can I use the T flip-flop of an IOB in VIRTEX devices ? Records #6214 and #7290 in the Xilinx Answers Database don't fix my problem. Thanks in advance to all who can help. Fred Laboratoire Signaux et Système du CNAM PARIS Sent via Deja.com http://www.deja.com/ Before you buy.Article: 21253
Hallo and a Good Day, I have the following solution (?) for a problem, but I cannot estimate its implications. Ok, there is a registered signal "withinReadCycle" that is asserted at the start of a read cycle , i.e. by a certain condition (see below "condition1") sunchronously to clock edge. The signal should be left asserted as long as another condition ("condition2") does not occure. But when the condition occures I would like to de-assert the signal "withinReadCycle" as quick as possible, that is I don't want to wait for the rising clock edge - which certainly means that I asynchronously de-assert that signal. I would code that as follows in VHDL: read_regs: PROCESS (CLKxCI, RSTxRLI) BEGIN -- PROCESS read_regs IF RSTxRLI = '0' THEN -- asynchronous reset (active low) withinReadCycle <= '0'; ELSIF (condition2 = '1') THEN withinReadCycle <= '0'; ELSIF CLKxCI'event AND CLKxCI = '1' THEN -- rising clock edge IF ( condition1 = '1' OR withinReadCycle = '1' ) THEN withinReadCycle = '1'; bla; ELSIF ( condition1 = '0' ) THEN withinReadCycle <= '0'; END IF; END IF; END PROCESS read_regs; And now, my questions: (1) Is there a chance to be able to synthesize that code? (2) The device for which to synthesize must have a product-term-driven reset, right? (3) In case that (1) can be answered with yes, would there be success for a Xilinx4036? I would appreciate any hint or comment and ... .. thank, you so far GuidoArticle: 21254
I try to generate a pipelined multiplier, but have the following error, any idea of the problem ? ERROR: An internal error has occurred. Please call Xilinx support. 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 mul16 WARNING: Core mul16 did not generate product ImpNetlist. WARNING: Warnings and/or errors encountered while generating mul16 (Variable_Parallel_Multiplier 1.0) All output products requested may not have been generated. Thanks, StevenArticle: 21255
To save the bother of installing the whole M1.5 series CDs again does anyone know which archive the 4020XL series of device personalities is held in,or even better,where I can get those mischievous 5 files from? Thanks in advance, Robert.Article: 21256
Has anyone ever tried to program a JTAG device (eg XC9500) using the PC parallel port directly connected to the JTAG pins? Got links to software routines? thanks, MH. Sent via Deja.com http://www.deja.com/ Before you buy.Article: 21257
Can any one tell me how to write a VHDL code to read bits (1s and 0s) from a file on the PC and another VHDL code to write output bits to a file on the PC? I am using both Altera and Xilinx platforms. Is it actually possible on these platforms ? Thanks in advance, Avinash Maddy maddy@eng.fsu.eduArticle: 21258
This is a multi-part message in MIME format. --------------536BEF03B9142CF0CD7A4F0F Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit All our XS Boards use the parallel port for programming, including the XC9500-based boards which use JTAG. We have our low-level programming routines at http://www.xess.com/xstools-source-2_0.zip and documentation at http://www.xess.com/xstools-src-doc-2_0.pdf. mark_harvey@my-deja.com wrote: > Has anyone ever tried to program a JTAG device (eg XC9500) using the PC > parallel port directly connected to the JTAG pins? Got links to software > routines? > > thanks, > MH. > > Sent via Deja.com http://www.deja.com/ > Before you buy. -- || Dr. Dave Van den Bout XESS Corp. (919) 387-0076 || || devb@xess.com 2608 Sweetgum Dr. (800) 549-9377 || || http://www.xess.com Apex, NC 27502 USA FAX:(919) 387-1302 || --------------536BEF03B9142CF0CD7A4F0F Content-Type: text/x-vcard; charset=us-ascii; name="devb.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Dave Vanden Bout Content-Disposition: attachment; filename="devb.vcf" begin:vcard n:Vanden Bout;Dave tel;fax:(919) 387-1302 tel;work:(919) 387-0076 x-mozilla-html:FALSE url:http://www.xess.com org:XESS Corp. adr:;;2608 Sweetgum Drive;Apex;NC;27502;USA version:2.1 email;internet:devb@xess.com title:FPGA Product Manager x-mozilla-cpt:;-16464 fn:Dave Vanden Bout end:vcard --------------536BEF03B9142CF0CD7A4F0F--Article: 21259
Juan Antonio Gómez Pulido <jangomez@unex.es> wrote: : Please, can anybody tell me if exits a PC simulator of parallel : processors? There are a number of varying types - which type you want depends on what you're trying to do: VHDL simulators: * 27000 Leapfrog VHDL Simulator * ModelSim - a VHDL simulator - http://www.model.com/ * MyVHDL - from Mycad - http://www.mycad.com/product/vhdl.html * VHDL System Simulator - from Synopsis * Cordis VHDL simulator Cellular automata simulators: * WebsideCA - The Isle Ex CA Explorer (Applet) - http://jmge.net/wca.htm * Mirek's Java Cellebration: http://www.mirwoj.opus.chelm.pl/mjcell/mjcell.html Alife parallel processing simulations: * AVIDA: http://www.krl.caltech.edu/avida/ -- __________ |im |yler The Mandala Centre http://www.mandala.co.uk/ tt@cryogen.com Some things have to be believed to be seen.Article: 21260
Hi all !! I need informations or references or the both on DCT implementation using FPGA !! If you've got it, please contact me !! Thks SEB -- . , '. '. \ \ ._ '-.'. `\ \ '-._; .'; `-.'. `~-.; '. '. '--,` '. -='. ; .--=~~=-, -.; ; .-=`; `~,_.; / ` ,-`' .-; | .-~`. .; ; Seb@stien.bizland.com .;.- .-; ,\ `.' ,=; .-' `~.-._ .'; .'; .' .' '-. .\ ; ; ,.' _ a', .'~";-` ; ;"~` `'-=.) .' .' . _; ;', ; '-.._`~`.' \ ; ; : `~' _'\\_ '\\_Article: 21261
What makes you think there is a toggle flip-flop in the Virtex IOB? Fig 2 on page 3-6 of the Xilinx data book clearly describes the three flip-flops as D-type. So you need to build the toggle logic in a CLB, but you can feed the state of the output back through the input logic. This gives fastest clock-to-out. Otherwise do the whole thing in a CLB and use the IOB just as buffer. Peter Alfke bfredc@my-deja.com wrote: > Hi. > > Would anyone know how can I use the T flip-flop > of an IOB in VIRTEX devices ? > Records #6214 and #7290 in the Xilinx Answers > Database don't fix my problem. > > Thanks in advance to all who can help. > > Fred > Laboratoire Signaux et Système du CNAM PARIS > > Sent via Deja.com http://www.deja.com/ > Before you buy.Article: 21262
Sorry, that was a dumb posting of mine. I saw "T-flip-flop" and did not realize that it, in this case, meant the flip-flop controlling the 3-state ( or output Enable ). Forget my toggle comments. Peter Alfke Peter Alfke wrote: > What makes you think there is a toggle flip-flop in the Virtex IOB? > Fig 2 on page 3-6 of the Xilinx data book clearly describes the > three flip-flops as D-type. > So you need to build the toggle logic in a CLB, but you can feed the > state of the output back through the input logic. This gives fastest > clock-to-out. Otherwise do the whole thing in a CLB and use the IOB > just as buffer. > > Peter Alfke > > bfredc@my-deja.com wrote: > > > Hi. > > > > Would anyone know how can I use the T flip-flop > > of an IOB in VIRTEX devices ? > > Records #6214 and #7290 in the Xilinx Answers > > Database don't fix my problem. > > > > Thanks in advance to all who can help. > > > > Fred > > Laboratoire Signaux et Système du CNAM PARIS > > > > Sent via Deja.com http://www.deja.com/ > > Before you buy.Article: 21263
All, Below is a link to a doc describing how we want to program multiple FPGAs on mulitple boards across a backplane in a fairly large system. Please review and comment. The most non-standard thing is not hooking up the INIT signals, but holding it low for a defined amount of time to ensure the FPGAs are ready. This way, we don't have to deal with a bidirectional signal on the backplane. Any and all comments welcome. http://www.arl.wustl.edu/burst/reprogram/reprogram.pdf TomArticle: 21264
Avinash Maddy wrote: > > Can any one tell me how to write a VHDL code to read bits (1s and 0s) from a > file on the PC and another VHDL code to write output bits to a file on the > PC? > I am using both Altera and Xilinx platforms. Is it actually possible on > these platforms ? > Thanks in advance, > Avinash Maddy > maddy@eng.fsu.edu I am pretty sure that this can be done, but it is not straightforward as it would be in C. You might want to post this message to comp.lang.vhdl. I would bet that you could get someone there to provide some sample code. -- Rick Collins rick.collins@XYarius.com remove the XY to email me. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design Arius 4 King Ave Frederick, MD 21701-3110 301-682-7772 Voice 301-682-7666 FAX Internet URL http://www.arius.comArticle: 21265
In your design, just use a regular D flip-flop. Don't put any combinatorial logic in front of it (this isn't absolutely necessary, but it makes it easier to meet timing) and don't drive anything other than the one tristate control with it. When compiling the design turn on the use-IOB flip-flops option in the mapper. That will push the flip-flop into the IOB as long as there is nothing goofy with it. bfredc@my-deja.com wrote: > Hi. > > Would anyone know how can I use the T flip-flop > of an IOB in VIRTEX devices ? > Records #6214 and #7290 in the Xilinx Answers > Database don't fix my problem. > > Thanks in advance to all who can help. > > Fred > Laboratoire Signaux et Système du CNAM PARIS > > Sent via Deja.com http://www.deja.com/ > Before you buy. -- -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: 21266
Ray Andraka wrote: > > In your design, just use a regular D flip-flop. Don't put any > combinatorial logic in front of it (this isn't absolutely necessary, but > it makes it easier to meet timing) and don't drive anything other than > the one tristate control with it. When compiling the design turn on the > use-IOB flip-flops option in the mapper. That will push the flip-flop > into the IOB as long as there is nothing goofy with it. Ok, I give up. How do I know if the FF is "goofy". ;-) -- Rick Collins rick.collins@XYarius.com remove the XY to email me. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design Arius 4 King Ave Frederick, MD 21701-3110 301-682-7772 Voice 301-682-7666 FAX Internet URL http://www.arius.comArticle: 21267
Tom McLaughlin wrote: > > All, > Below is a link to a doc describing how we want to program multiple > FPGAs on mulitple boards across a backplane in a fairly large system. > Please review and comment. The most non-standard thing is not hooking > up the INIT signals, but holding it low for a defined amount of time to > ensure the FPGAs are ready. This way, we don't have to deal with a > bidirectional signal on the backplane. Any and all comments welcome. > > http://www.arl.wustl.edu/burst/reprogram/reprogram.pdf > > Tom I don't see anything wrong with the way you are using the INIT, DONE and other signals. I don't completly understand why you are driving two clock lines from the misc card. Can't you just drive all the clock lines from the Master card? Actually it just occured to me why you might be driving the clock from the Misc board. If you are worried about the delays, then the signal out from the PROM would have the least skew from the clock this way. One point that I think might be significant is the circuit you use to delay the INIT signal to the PROM. You will need a delay in the 100s of uS or mS range IIRC. A FF delay circuit as you have shown will take many, many FFs unless you use a very slow clock. You could do better by using a counter which is reset when the incoming INIT signal is at zero. When INIT goes high, the counter starts counting until it reaches max count. That signal is used to disable the counting and assert the INIT to the PROM. Reset has to have priority over the count enable. This will give you a much longer period with many fewer FFs. -- Rick Collins rick.collins@XYarius.com remove the XY to email me. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design Arius 4 King Ave Frederick, MD 21701-3110 301-682-7772 Voice 301-682-7666 FAX Internet URL http://www.arius.comArticle: 21268
--------------F22A3C3FBA5E4F438BDC1C0F Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; x-mac-creator="4D4F5353" Content-Transfer-Encoding: 7bit >>From my computer here at Xilinx, I can access all sorts of semiconductor websites. It's a joy. I have Intel, AMD, National, Altera, Quicklogic, Actel, Cyprus at my fingertips. But not Atmel. I get: Forbidden You don't have permission to access / on this server. Apache/1.3.9 Server at www.atmel.com Port 80 Shame on you, Atmel ! What are you afraid of ? Peter Alfke --------------F22A3C3FBA5E4F438BDC1C0F Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit <!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html> >>From my computer here at Xilinx, I can access all sorts of semiconductor websites. <br>It's a joy. <br>I have Intel, AMD, National, Altera, Quicklogic, Actel, Cyprus at my fingertips. <br>But not Atmel. <br>I get: <h2> <b><font color="#FF0A37">Forbidden</font></b></h2> <h2> <b><font color="#FF0A37">You don't have permission to access / on this server.</font></b></h2> <h2> <b><font color="#FF0A37">Apache/1.3.9 Server at www.atmel.com Port 80</font></b></h2> <font color="#000000">Shame on you, Atmel ! What are you afraid of ?</font><b><font color="#000000"></font></b> <p><font color="#000000">Peter Alfke</font></html> --------------F22A3C3FBA5E4F438BDC1C0F--Article: 21269
Rickman wrote: > One point that I think might be significant is the circuit you use to > delay the INIT signal to the PROM. You will need a delay in the 100s of > uS or mS range IIRC. A FF delay circuit as you have shown will take > many, many FFs unless you use a very slow clock. You could do better by > using a counter which is reset when the incoming INIT signal is at zero. > When INIT goes high, the counter starts counting until it reaches max > count. That signal is used to disable the counting and assert the INIT > to the PROM. Reset has to have priority over the count enable. This will > give you a much longer period with many fewer FFs. I can't actually view the proposal document at the moment, because our proxy server is down, but if this cdelay ircuit is external to any programmable logic device, I'd suggest you just use a one-shot (such as a 74121) for this purpose. Why does everyone thing that the best solution to everything must be digital these days?! ;-) MarkArticle: 21270
I am fairly sure that this design will work, but I would have designed it differently. You already have one special board containing the data source, the PROMs. I would have made this the self-contained configuration source, controlled by a small CPLD and a crystal oscillator ( total <$ 10.- ) Advantages: CCLK is unidirectional and can be properly terminated, and definitely not series-terminated ( I see problems with the series termination in the present design, generating half-amplitude signals). Configuration will be faster, since the oscillator is crystal-controlled. Can cut configuration time in half. The CPLD can also generate the stretch-time for INIT, with any desired value. I would also include a precision voltage monitor chip ( $ 1 ) to start counting when Vcc is definitely there. Just my $ 0.02 worth Peter Alfke, Xilinx Applications Tom McLaughlin wrote: > All, > Below is a link to a doc describing how we want to program multiple > FPGAs on mulitple boards across a backplane in a fairly large system. > Please review and comment. The most non-standard thing is not hooking > up the INIT signals, but holding it low for a defined amount of time to > ensure the FPGAs are ready. This way, we don't have to deal with a > bidirectional signal on the backplane. Any and all comments welcome. > > http://www.arl.wustl.edu/burst/reprogram/reprogram.pdf > > TomArticle: 21271
Well, the good news is that Boulder can still access the sight. You can refer information needs to us, and we can pass it on to the rest of the Xilinx community.Article: 21272
Peter Alfke (peter@xilinx.com) posted duplicate MIME-encoded copies of: : From my computer here at Xilinx, I can access all sorts of : semiconductor websites. : It's a joy. : I have Intel, AMD, National, Altera, Quicklogic, Actel, Cyprus at my ^^^^^^ Cypress? : fingertips. But not Atmel. I get: : Forbidden : You don't have permission to access / on this server. : Apache/1.3.9 Server at www.atmel.com Port 80 : Shame on you, Atmel ! What are you afraid of ? So look at them from home, or use a redirector service like www.anonymiser.com. - Larry Doolittle <LRDoolittle@lbl.gov>Article: 21273
OK, so I've got this design that's supposed to run at 80 MHz, and it fits into about 60% of a 4013XLA. Now, I'd read an appnote that mentioned the Fast Capture Latches, and how they're faster than using just the input flop in the IOB. So, I instantiated all of these things, and because the early clocks only go to part of the chip, I used three clock pins (all driven by three different outputs of a PLL-type clock buffer). Now, since the fast capture latch requires both an early clock and a global-low-skew clock, I had to instantiate them both. Place, route, meets timing. I noticed something weird in the map report. For each clock pin, a BUFG, a BUFGLS and a BUFGE were used. Looking at it with the FPGA Editor, I noticed that indeed, the clock pin fed the BUFG, which fed both the BUFGLS and the BUFGE. I don't know what this means. FPGA Express didn't stick the BUFG there, as far as I can tell. I then ripped out all of the fast capture latches and just used a single clock. Met timing fine. And the only clock buffer inferred was the BUFGLS, which is what one would expect. -- a ----------------------------------------- Andy Peters Sr Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters (at) noao \dot\ edu "Money is property; it is not speech." -- Justice John Paul StevensArticle: 21274
Rickman wrote in message <38CEA92E.E3A21A45@yahoo.com>... >Ray Andraka wrote: >> >> In your design, just use a regular D flip-flop. Don't put any >> combinatorial logic in front of it (this isn't absolutely necessary, but >> it makes it easier to meet timing) and don't drive anything other than >> the one tristate control with it. When compiling the design turn on the >> use-IOB flip-flops option in the mapper. That will push the flip-flop >> into the IOB as long as there is nothing goofy with it. > >Ok, I give up. How do I know if the FF is "goofy". ;-) They get that way when you slip them a Mickey. -- a ----------------------------------------- Andy Peters Sr Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters (at) noao \dot\ edu "Money is property; it is not speech." -- Justice John Paul Stevens Reply-To: "Sherdyn" <sherdyn@yahoo.com>
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