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
The brute force way (simplest design perhaps) would be to build an 8-bit lookup table, and just record the number of ones for each entry (0-255)..so it would be a 256x8 ROM. Use it 3 times for 1 24-bit word...Pretty fast too.Article: 20251
Björn Lindegren wrote in message <7qYl4.4016$jg4.8056@nntpserver.swip.net>... >Hi. >Using Xilinx Foundation series 2.1i, and I want to build a program with tree >states to XC9536 in the programming language Abel via State Editor. > >In state number one outport q1 activates and all the others are deactivated. >In state no two outport q2 activate and q1, q2 is activated and in the last >state is only q3 activated. > >When I program XC9536 with the code in the end of this message, all outports >(q1, q2, q3) are activated. > >What's wrong? > >Thankful for help > >Björn Lindegren >" >" File: E:\PROGRAM\XFS2.1\ACTIVE\PROJECTS\95360130\delay.abl >" created: 02/02/00 16:31:30 >" from: 'E:\PROGRAM\XFS2.1\ACTIVE\PROJECTS\95360130\delay.asf' >" by: fsm2hdl - version: 2.0.1.53 >" >module delay >Title 'delay' > >Declarations > >"clocks >CLK PIN 43; > >"input ports > >"output ports >q1 PIN 22; >q2 PIN 21; >q3 PIN 13; > >"******** SYMBOLIC state machine: Sreg0 ****** >Sreg0 STATE_REGISTER; >S1, S2, S3 STATE; > > >"diagram ACTIONS > >"************* state machine: Sreg0 ************* > >Equations >" clock signals definitions > Sreg0.clk = CLK; > >State_diagram Sreg0 > > >State S1: > > q1=1; > q2=0; > q3=0; > x=x+1; > IF ('x<1000') THEN > S1 > ELSE IF ('x=1000') THEN > S2; > > >State S2: > > q1=0; > q2=1; > q3=0; > x=x+1; > IF ('x<2000') THEN > S2 > ELSE IF ('x=2000') THEN > S3; > > >State S3: > > q1=0; > q2=0; > q3=1; > x=x+1; > IF ('x<3000') THEN > S3 > ELSE IF ('x=3000') THEN > S1 WITH > x=1; > ENDWITH; > >" end of state machine - Sreg0 > > >end delay You don't have a reset state or a default state. Sreg starts up in an unknown state and can't transition to a known state. -- a ----------------------------------------- Andy Peters Sr Electrical Engineer National Optical Astronomy Observatories 950 N Cherry Ave Tucson, AZ 85719 apeters (at) noao \dot\ edu Spelling Counts! You don't loose your money - you lose it.Article: 20252
This is a multi-part message in MIME format. --------------D49C00D5DF504C438744CE85 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit "Björn Lindegren" wrote: > Hi. > Using Xilinx Foundation series 2.1i, and I want to build a program with tree > states to XC9536 in the programming language Abel via State Editor. > > In state number one outport q1 activates and all the others are deactivated. > In state no two outport q2 activate and q1, q2 is activated and in the last > state is only q3 activated. > > When I program XC9536 with the code in the end of this message, all outports > (q1, q2, q3) are activated. > > What's wrong? In earlier versions of the State editor (in Foundation 1.5), I noticed a similar problem. Outputs would not work correctly if you used vector equations such as "q = ^b010". You had to explicitly split-up the equation into three equations: "q0=0; q1=1; q2=0;" and attach those to the output states. Then the resulting design would work. --------------D49C00D5DF504C438744CE85 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;David 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:;28560 fn:Dave Vanden Bout end:vcard --------------D49C00D5DF504C438744CE85--Article: 20253
Steve, do you have the latest Service Pack (which is SP4). Xilinx had a DLL problem in Service Pack 3 (or 2? I dont recall). Marco ---------------------------------------------------------------------------- Marco Winzker Liesegang electronics (as always, speaking only for myself)Article: 20254
If you want to go real "brute force" and fast, use the Virtex or Virtex-E BlockRAM with two levels: First level four 4k x 1 BlockRAMs. They are dual-ported, so each can look at 12 inputs on one port, plus the other 12 inputs on the other port, each port generating 1 output per BlockRAM, for a total of 8 outputs. ( The look-up code is the same for both ports !) Second level: 256 x 16, converting the 8 signals into the appropriate code, could of course be only 5 bits. BlockRAM has synchronous read, so you have one extra clock period of latency, but the clock could run at over 150 MHz. If you don't like the latency, build the 4-bit adder in CLBs. The "price" of BlockRAMs is a bit strange: If you have barely enough, any additional BlockRAM would be very expensive. If you don't use them all, the remaining BlockRAMs are free, and they can be used ( or abused ) as shown above. Peter Alfke, Xilinx Applications ======================= Nick Macias wrote: > The brute force way (simplest design perhaps) would be to build an 8-bit > lookup table, and just record the number of ones for each entry > (0-255)..so it would be a 256x8 ROM. Use it 3 times for 1 24-bit > word...Pretty fast too.Article: 20255
Tom Burgess (tom.burgess@hia.nrc.ca) wrote: : EDN mag, Jan 6th issue, p. 24 had some unattributed : Spartan-2 price projections: ^^^^^^^^^^^ : XC2S15 3.95, XC2S30 4.95, XC2S50 7.95, XC2S100 9.95, XC2S150 12.95 : The fine print states that this pricing is for 250,000 units, lowest : speed and cheapest package, end of 2000. They'll be throwing in a free : pizza and beverage with every million-dollar order, too :) Thanks. And I guess I shouldn't have been fooled into thinking that production in 1Q00 meant anything sooner than March 31. My guesses for small quantity prices at that time stands. As someone pointed out, the first price break comes at 25 (for Spartan, at least), and counts for a good 10-20%. I put up a web page to make searches for _real_ P&A easier: http://recycle.lbl.gov/xc2search.html Still nothing :-( - Larry Doolittle <LRDoolittle@lbl.gov>Article: 20256
Hello, Is there a way to download a Foundation software for free somewhere? Because I am from Serbia there is no way to pay for it. If anybody have it and want to send me somehow please notify me by email. I have 3 Xilinx devices but I can't configure them. I would be grateful if someone help me. veselic@eunet.yu Veselic MladenArticle: 20257
Hi, I have recently started working with Xilinx Foundation tools and I am still on the learning curve with VHDL. My previous experience is AHDL for Altera. I am having troubles understanding what synthesis does with the code I write so I would like to be able to see it somehow in the schematic or similar format. I know that Sinplify has this option but I am not going to have it in the hear future. Is there any way to do this apart from manual reconstructing schematics from the EDIF netlist? I will appreciate any input on this matter. Thanks. Mikhail MatusovArticle: 20258
Does any place in the UK sell Altera devices to the hobbyist.Article: 20259
Are you Serbian? If yes, have you been a naughty boy? If yes, then no. Veselic Mladen <veselic@eunet.yu> wrote in message news:87a83l$2k$1@SOLAIR2.EUnet.yu... > Hello, > > Is there a way to download a Foundation software for free somewhere? > Because I am from Serbia there is no way to pay for it. > If anybody have it and want to send me somehow please notify me by email. > I have 3 Xilinx devices but I can't configure them. > > I would be grateful if someone help me. > > veselic@eunet.yu > > Veselic Mladen > >Article: 20260
geesh! Us poor sods have to work full time. You youngsters don't know wot hardwork is. Jamil Khaib <Khatib@ieee.org> wrote in message news:3896CE3C.96E675E0@ieee.org... > Hi > is there any part time job that I can do via the internet > > I know VHDL and FPGA design and basics of verifications > > Thanks >Article: 20261
Hello, I am looking for a CPU core which can be placed in an FPGA. It should have the following features: 1. 32-bit registers and ALU, integer only; don't need multiply or divide. 2. Fast enough to run at 66 MHz on a Virtex or Apex FPGA. 3. Not too big (a single instance should fit in less than 100K system gates, whatever it means). 4. Possible to have more than one instance in a single chip; not locked to specific cells or I/O pins. 5. Must have code development tools (assembler, linker, debugger) available; C compiler is nice-to-have but not mandatory. 6. Preferably synthesizable VHDL or Verilog; if available as netlist or routed block, must have a VHDL simulation model. 7. Can be converted later to an ASIC cell. I would greatly appreciate any pointers; after all replies (if any) have been sent, I will post a summary in the relevant NGs. Please send replies to my email (assaf_sarfati@yahoo.com) as well as posting them; I suspect my NG server either loses posts or deletes them after a few minutes. Thanks in Advance Assaf SarfatiArticle: 20262
funky jim <jmc8197@hotmail.com> wrote: : Does any place in the UK sell Altera devices to the hobbyist. To quote from http://ftp.altera.com/html/office.6.html UNITED KINGDOM: Ambar-Cascom, Ltd. The Gatehouse Gatehouse Way Aylesbury, Buckinghamshire HP19 3DL United Kingdom Flexible Logic (Arrow) St.Martins Business Centre Cambridge Road Bedford MK42 0LF United Kingdom Thame Components, Ltd. Thame Park Road Thame Oxfordshire OX9 3UQ United Kingdom URL: http://www.tcl.memec.com/ [I snipped out phone numbers and email addresses so I'm less responsible for any spam these folks get. See the original page for these.] The URL at the end has lots of Altera bits and pieces - but they don't seem to be stuck together very much. http://www.computer-solutions.co.uk/ also have Altera parts, I believe. This information is probably of low quality - but it's the best I can manage :-| -- __________ |im |yler The Mandala Centre http://www.mandala.co.uk/ tt@cryogen.com Never attempt to replicate a successful experiment.Article: 20263
The only Windows calls that are being translated are some file IO at the beginning and end of each program, 99.9999% of the time is spent running straight C code with no OS calls. There shouldn't be any performance difference at all. Josh eml@riverside-machines.com.NOSPAM wrote: > > On Sun, 30 Jan 2000 20:06:15 -0500, "B. Joshua Rosen" > <bjrosen@polybus.com> wrote: > > >I've just put up a HowTo page for running Xilinx place and route tools under Linux. > > > >http://www.polybus.com/xilinx_on_linux.html > > Very nice work, but what about the overhead of running the windows > emulator? Is this significant? > > Thanks > > EvanArticle: 20264
Uh, sign me up too... I've got time, tools, and smarts. m Matt Billenstein http://w3.one.net/~mbillens/ mbillens@one.net "Jamil Khaib" <Khatib@ieee.org> wrote in message news:3896CE3C.96E675E0@ieee.org... | Hi | is there any part time job that I can do via the internet | | I know VHDL and FPGA design and basics of verifications | | Thanks |Article: 20265
On Wed, 02 Feb 2000 22:18:17 GMT, "Mikhail Matusov" <matusov@ANNTIsquarepegSPPAMM.ca> wrote: >Hi, > >I have recently started working with Xilinx Foundation tools and I am still >on the learning curve with VHDL. My previous experience is AHDL for Altera. > >I am having troubles understanding what synthesis does with the code I write >so I would like to be able to see it somehow in the schematic or similar >format. I know that Sinplify has this option but I am not going to have it >in the hear future. Is there any way to do this apart from manual >reconstructing schematics from the EDIF netlist? > >I will appreciate any input on this matter. Thanks. > >Mikhail Matusov Any reason why you can't get an evaluation copy of synplify? The guy sitting at the desk next to mine just got one, and seems quite impressed by the schematic viewer. Regards, Allan.Article: 20266
I played around, and got this 24-bit design down to three BlockRAMs plus one CLB, in Virtex or Spartan-II. The design can be clocked at >150 MHz. It uses the first clock edge to clock in the 24 bits of data, then uses the next clock to generate the binary result, i.e it is pipelined and thus can generate a new result every clock. Here are the details: Use 22 bits to dual-port address two 2k x 2 BlockRAMs. Each BlockRAM has two outputs per port, for a total of eight outputs, in two sets of 4-bit binary-weighted signals. Remember that the BlockRAM, used as a ROM here, is a clocked device. Feed the remaining two input bits to one LUTs, generating only the sum (ignore the carry). Use the 8 BlockRAM outputs and the two remaining inputs to address one 1k x 4 BlockRAM which generates the 4 MSBs of the 5-bit result. Generate the LSB of the result in one 3-input LUT, in parallel with the BlockRAM. All LUTs have their own pipelining flip-flop "for free". Not too expensive and very fast, if you can tolerate the pipelining. Thanks for the brain-stretcher ! Peter Alfke, Xilinx Applications Peter Alfke wrote: > If you want to go real "brute force" and fast, use the Virtex or Virtex-E > BlockRAM with two levels: > > First level four 4k x 1 BlockRAMs. They are dual-ported, so each can look > at 12 inputs on one port, plus the other 12 inputs on the other port, each > port generating 1 output per BlockRAM, for a total of 8 outputs. ( The > look-up code is the same for both ports !) > Second level: 256 x 16, converting the 8 signals into the appropriate code, > could of course be only 5 bits. > BlockRAM has synchronous read, so you have one extra clock period of > latency, but the clock could run at over 150 MHz. > If you don't like the latency, build the 4-bit adder in CLBs. > > The "price" of BlockRAMs is a bit strange: If you have barely enough, any > additional BlockRAM would be very expensive. If you don't use them all, the > remaining BlockRAMs are free, and they can be used ( or abused ) as shown > above. > > Peter Alfke, Xilinx Applications > ======================= > > Nick Macias wrote: > > > The brute force way (simplest design perhaps) would be to build an 8-bit > > lookup table, and just record the number of ones for each entry > > (0-255)..so it would be a 256x8 ROM. Use it 3 times for 1 24-bit > > word...Pretty fast too.Article: 20267
Hi, I asked this question at sci.electronics.design already and was sent here: I'm looking for the best solution of crossing clock domain boundaries in digital designs solving the related timing problems. In scan pathes lock-up latches are common (like Synopsys inserts them during scan insertion). Why negative (level triggered) latches instead of negative (edge triggered) FFs? Is the setup/hold problem at the FALLING clock edge the reason. which I do not have with latches (according to my simulations) ? Couldn't I use them in functional pathes too? Should I more concentrate on the adjustment of the clocks of the different modules in my design or more on the datapathes. It seems easy if the datapath is unidirectional but if it goes in both directions? More detailed: At the moment I have clocks which have a common frequency base (like 32 Mhz and 64 Mhz). The phase relation is at least not yet predictable as the clocks are driven by seperate clocktrees (for each frequency) which we not yet have inserted. I look for a solution to avoid problems with clock adjustment after layout (clock tree insertion) and the timing on pathes that originate in one domain and end at the other. Any comments, help or hints for information sources are very welcome. VolkerArticle: 20268
Please help me, I have designed a FIFO core for FPGA using schematic and VHDL. I got good results using the schematics but the VHDL before synthesis I made many archetectures and simulated all of them they gave me fine results but after synthesizing them I got very bad results. After synthesis, I got problem with the generation of Flages which control the read and write operation. through simulationthe flages do not change when an event is trigered they can only be reseted the simulator shows 'U' state all time except after reset. why do I have this problem is there any thing with my VHDL code? Note: I tried to synthesize the code on many synthesizers and targeted to many PLD devices but all gave me the same results Can you please help me My main FIFO core page http://www.geocities.com/SiliconValley/Pines/6639/ip/fifo.html The VHDL code which contains the 3 archs I synthesized FIFO_v4, FIFO_v5, FIFO_v6 http://www.geocities.com/SiliconValley/Pines/6639/ip/fifo_vhdl.html These are the schematic results of the synthesized code generated by siniplify http://www.geocities.com/SiliconValley/Pines/6639/ip/fifo_v4.pdf http://www.geocities.com/SiliconValley/Pines/6639/ip/fifo_v5.pdf http://www.geocities.com/SiliconValley/Pines/6639/ip/fifo_v6.pdf This is the wave form of the synthesized code http://www.geocities.com/SiliconValley/Pines/6639/ip/wave.jpg The FIFO project uses the idea of OpenIPCore desgin methodology You can get the VHDL code and the test vectors from OpenCores CVS Thanks in advance Jamil Khatib OpenIP Organization http://www.openip.org OpenIPCore Project http://www.openip.org/oc OpenCores Project http://www.opencores.orgArticle: 20269
hello, I build the evaluationboard for Xilinx XC9536 (http://www.xilinx.com/xapp/xapp078.pdf) and the JTAG-cable (http://www.xilinx.com/support/programr/jtag_cable.pdf) but it won't work. every time I try to download a jed-file, the target answered the component-ID : 11111111000000000000000000000011 and this seems not to be a Xilinx-ID (second bit must be zero). I checked the hardware several time, but I didn't found an error. -> is there a simple solution which is not commented in application notes? Wilken Klee LOG-file: JTAG Programmer Started 2000/02/03 09:33:50 Loading Boundary-Scan Description Language (BSDL) file 'c:/progs/xilinx/cpld/xc9500/data/xc9536.bsd'.....completed successfully. Checking boundary-scan chain integrity...done. Verifying device positions in boundary-scan chain... Instance 'eval(Device1)' at position '1'...ERROR:JTag - The IDCODE returned (11111111000000000000000000000011) indicates that the instance 'eval(Device1)' in the boundary-scan chain is not a Xilinx part. failed. Verification completed. ERROR:JTag - Boundary scan chain has been improperly specified. Please check your configuration and re-enter the boundary-scan chain information. Boundary-scan chain validated unsuccessfully. ERROR:JTag - : The boundary-scan chain has not been declared correctly. Verify the syntax and correctness of the device BSDL files, correct the files, reset the cable and retry this command.Article: 20270
"Allan Herriman" <allan.herriman.hates.spam@fujitsu.com.au> a écrit dans le message news: 3899167f.8127266@newshost.fujitsu.com.au... > On Wed, 02 Feb 2000 22:18:17 GMT, "Mikhail Matusov" > <matusov@ANNTIsquarepegSPPAMM.ca> wrote: > > >Hi, > > > >I have recently started working with Xilinx Foundation tools and I am still > >on the learning curve with VHDL. My previous experience is AHDL for Altera. > > > >I am having troubles understanding what synthesis does with the code I write > >so I would like to be able to see it somehow in the schematic or similar > >format. I know that Sinplify has this option but I am not going to have it > >in the hear future. Is there any way to do this apart from manual > >reconstructing schematics from the EDIF netlist? > > > >I will appreciate any input on this matter. Thanks. > > > >Mikhail Matusov > > Any reason why you can't get an evaluation copy of synplify? > The guy sitting at the desk next to mine just got one, and seems quite > impressed by the schematic viewer. > > Regards, > Allan. You have a schematic viewer in FPGA Express too (so in Fondation Express).Article: 20271
Hi I am trying to implement a design in a Spartan 2 device. I (think I) updated Foundation to enable these devices but I still don't manage to do it. When I want to synthesize (with FPGA Express) my design, I can not choose the Spartan2 family (the family is not in the list). A Xilinx FAE told me to choose Virtex family for synthesis and then Spartan2 for mapping and P&R but I don't know where to do this. If anyone can help... Thanks in advance Nicolas MATRINGE DotCom S.A. Conception electronique 16 rue du Moulin des Bruyeres Tel 00 33 1 46 67 51 11 92400 COURBEVOIE Fax 00 33 1 46 67 51 01 FRANCEArticle: 20272
I'm having a problem figuring out how to get a group of people to share a common set of setups for Renoir. (i) Engineer A works locally, and creates or modifies library A on machine A. (ii) Engineer B creates/modifies library B on machine B. (iii) A and B use a revision control system that archives onto server S. (iv) Engineer C now needs access to libraries A and B, and so checks them out of the archive. Problem: engineer C cannot use the libraries. The reason is that Renoir has 2 local initialisation and preferences files, which are different for libraries A and B. These files contain information on library usage and changes. To fix this problem, all the engineers have to use a common copy of one initialisation and one profile file, referencing all the libraries, with the masters on the server. However, this doesn't work well. If engineer C checks outs the initialisation files from the server, then they're locked, and no-one else can modify them. Engineers A and B can actually use Renoir at the same time but, when Renoir exits, it realises that the files are locked by C, and so it silently creates a new local copy of the files, which screws up the archiving, as well as potentially losing any changes. In short, there doesn't seem to be a way for multiple engineers to use a common set of libraries, which is pretty annoying if you're bought several licences. One fix might be to allow each engineer to have access to only one library at a time, with each library having its own initialisation files, but then you have to mess around with environment variables to point to the required initialisation files. Has anyone come across this? Thanks EvanArticle: 20273
In article <38995b2c.93043005@news.dial.pipex.com>, <eml@riverside-machines.com.NOSPAM> wrote: >However, this doesn't work well. If engineer C checks outs the >initialisation files from the server, then they're locked, and no-one >else can modify them. Engineers A and B can actually use Renoir at the >same time but, when Renoir exits, it realises that the files are >locked by C, and so it silently creates a new local copy of the files, >which screws up the archiving, as well as potentially losing any >changes. Is this locking done by Renoir, or the version control tool? Is Renoir sensitive to the actions of the version control in any way? Most version control tools are highly unsuited to I.C. design; you need to find one that will let you share objects in a read-only manner.Article: 20274
> > You have a schematic viewer in FPGA Express too (so in Fondation Express). Could you please let me know how to enter the schematic viewer program in Foundation 2.1. Because I cannot find nothing like that. -- ----------- Ernest Jamro ------------ e-mail: jamro@uci.agh.edu.pl Address: AGH Technical University, Institute of Electronics Mickiewicza 30; 30-059 Kraków; Poland; tel. 48-12-617-2792
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