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 Tyler wrote: > = > 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 responsibl= e > 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 :-| One thing I would add to this is check prices before you decide what you're going to use. I was using EPM7096's which aren't isp compatible and which cost =A325 until I found I could buy the bigger EPM7128S which is isp = compatible for =A311. You can also get them through Farnell for about the= same price. Nial.Article: 20276
Hi All! What book do you recommend for beginners in VHDL Programming, which are using XILINX Foundation 2.1i. I am interested in books written in englisch or german language. Thank you for suggestions. -- Holger Kleinert Development / Support IBP Instruments GmbH Sutelstrasse 7a D-30659 Hannover, Germany http://www.ibpmt.com Fon : +49-511-651647 Fax : +49-511-652283Article: 20277
Hi, I haven't "Foundation 2.1." but ViewLogic CAO (Schematic, FPGA Express and so on) + Alliance 2.1 (Xilinx) With ViewLogic, it's an option to have schematic in FPGA Express. Ask your tech. support to know if it's not the same. (If you have schematic viewer, you normally right click in the "chip window" and "View Schematic", if you bought the option, it's work !) "Ernest Jamro" <jamro@uci.agh.edu.pl> a écrit dans le message news: 38998E08.A189F4E5@uci.agh.edu.pl... > > > > 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 > >Article: 20278
This is a multi-part message in MIME format. --------------0C23D22D8DD6CFF218436222 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Holger Kleinert wrote: > Hi All! > > What book do you recommend for beginners in VHDL Programming, which are > using XILINX Foundation 2.1i. > I am interested in books written in englisch or german language. > > Thank you for suggestions. > I'll give you my suggestions: "The Practical Xilinx Designer Lab Book" has many of the examples done with VHDL. That's a decent place to start if you are working specifically with the Xilinx Foundation tools. "VHDL for Programmable Logic" by Kevin Skahill is a good book. It is oriented toward Cypress devices and their Warp tools. Both of the books listed above discuss a subset of the VHDL language that's sufficient to get started. "The Designer's Guide to VHDL" by Ashenden is my bible for VHDL. It works very well as a reference. You can also check the "Related Books" link at http://www.optimagic.com for a list of VHDL books. Also check the "HDL Tutorials " link for online stuff you can get for free. --------------0C23D22D8DD6CFF218436222 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 --------------0C23D22D8DD6CFF218436222--Article: 20279
As they say, their recommendation was intended to cover a broad variety of applications. In your particular application, you might be able to get away with less, especially if you are not using many output drivers. You might want to look at the AVX SMT tantalum catalog (download pdf from http://www.avxcorp.com/products/capacitors/smtc.htm) for quantative info on ESR v.s. size for this line of caps. It looked like (as I recall) about 100 milliohms for 470 uF, v.s. 3-400 milliohms for the 47 uF. (at 100 KHz). For board layout, I would try to follow the recommendation, and then, if the application was cost sensitive, experiment with removing the big cap while measuring worst case power plane noise at the part. As for startup in your low-power app, you should (if you haven't already) read the "Power-On Power Supply Requirements" section of the data sheet (bottom of p. 24 in the DS003 v1.7 sheet). Essentially, it says that Virtex parts need lots of current (0.5 to 2A) during power-up, which is good to know early in the design process :) And don't forget the 50 ms max. supply ramp-up time. regards, tom John Janusson wrote: > > Hello: > > I have read XAPP158 (Powering Virtex FPGAs > http://www.xilinx.com/xapp/xapp158.pdf) and found the following > recommendations: > > VCCINT -> Guideline > ------------------- > 0.1 uF -> One per VCC > 47 uF -> Four per device (XCV50 - XCV300) > 470 uF -> One per device > > 470uF!!! In my experience, this seems excessive, even after reading the > disclaimer in Answers record #777 > (http://support.xilinx.com/techdocs/777.htm)... I was planning on following > the guidelines sans 470uF cap. It's a low power design in a XCV100 running > at mostly 4 to 32 MHz... > > Can others comment on power related issues in Virtex parts, particularly > related to decoupling and startup??? > > Thanks... > > John -- Tom Burgess -- Digital Engineer National Research Council of Canada Herzberg Institute of Astrophysics Dominion Radio Astrophysical Observatory P.O. Box 248, Penticton, B.C. Canada V2A 6K3 Email: tom.burgess@hia.nrc.ca Office: (250) 490-4360 Switch Board: (250) 493-2277 Fax: (250) 493-7767Article: 20280
only 10pF from TCK to GND and I can read the correct ID (and I can read signatur&user-code) but I still can't program this CPLD. It seems that my board is not routed fine. LOG-file: 'eval(Device1)': Programming device....ERROR:JTag - The program operation failed for instance 'eval(Device1)' when attempting to address location '0x12492'. The part appears to be in read or write protect mode. You will have erase the part the disable this mode. --> If the part is not in this mode then check that the target system power supply is stable, adequate and at the correct level. In addition make certain that each device has adequate decoupling. <--Article: 20281
Thanks to everybody for the input. I tried VISTA viewer in the FPGA Express and it works not bad at all. MMArticle: 20282
John Janusson wrote: > > Hello: > > I have read XAPP158 (Powering Virtex FPGAs > http://www.xilinx.com/xapp/xapp158.pdf) and found the following > recommendations: > > VCCINT -> Guideline > ------------------- > 0.1 uF -> One per VCC > 47 uF -> Four per device (XCV50 - XCV300) > 470 uF -> One per device > > 470uF!!! In my experience, this seems excessive, even after reading the > disclaimer in Answers record #777 > (http://support.xilinx.com/techdocs/777.htm)... I was planning on following > the guidelines sans 470uF cap. It's a low power design in a XCV100 running > at mostly 4 to 32 MHz... > > Can others comment on power related issues in Virtex parts, particularly > related to decoupling and startup??? > > Thanks... > > John This is the third time I have attempted to reply to this posting. The other two got lost in the ether. Maybe I should just keep my mouth shut? My opinion is that Xilinx is practicing some serious overkill without good justification. Their note about not knowing what the different designs are doing is real, but there are rules of physics that electronics follow. The series inductance of most capacitors other than ceramic chip caps is significant at any frequency above about 1 MHz. Some specify the ESR at 100 KHz. The series inductance of these caps is large enough that the size of the capacitor or the location or quantity of parts is not significant for typical switching noise which is in the range well above 100 MHz. So for switching noise, only ceramic chip caps located close to your power and ground pins will do a good job. Further these caps are inductive at the significant frequencies. The graph in the Xilinx app note show that the impedance above 100 MHz is not significantly different between same size parts with different capacitance. So it really doesn't make much difference if you use 0.1 uF or 0.01 uF caps. But certainly using 0.1 uF caps make you feel better, I know it does me :) But the cap size does make a real difference. Smaller is generally better, but wider makes more of a difference according to the manufacturer's data. So I would not recommend that you use multiple large capacitors. Rather use one large (and 470 uF is really large) value tantalum cap near the power connector for each supply voltage. Then use a 0.01 to 0.1 uF cap as close as possible to each power and ground pin pair. I ended up with some 90 pieces of 0.1 uF caps on my board giving me a total of about 4 - 5 uF per supply voltage in high freq decoupling. And it was a very small board, PC/104! The large capacitors are only good for lower freq signals. My 5 to 3 volt power converter only has 100 uF on the output for the 300 KHz switching freq. With that there is only about 20 - 30 mV of noise from a 1.5 Amp supply. So I don't think you need a lot more for chip decoupling. -- 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: 20283
On Thu, 03 Feb 2000 12:31:50 GMT, dej@coup.inode.org (David Jones) wrote: >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? The VC tool (MKS's Source Integrity) does the locking, and Renoir doesn't know anything about it, other than noticing the read-only attribute when it eventually tries to write the preferences file. In principle, Renoir could handle the version control itself. It does this by writing scripts for RCS, but I don't think that this would help. An individual user would still have to lock the global preferences file in order to get write access to it, and this would lock out the other users. In practice, Renoir is only used on a small part of the project anyway, and so it couldn't be used for VC since it doesn't know about the rest of the project. >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. I'm not sure that this is the problem - everyone can read the preferences file; it's just that Renoir's not intelligent enough to realise that multiple users might want to write to the preferences file. I think the answer might be that there's no way to have a global setup. Each designer has a local renoir.ini and .renoirprefs for each library, and these are archived together with that particular library. A user on another machine who needs the library checks it out together with its own ini and prefs file, and then runs Renoir with the -inifile and -prefsfile options pointing to the local files. This should give access to the local library and preserve the original designer's downstream tool settings, and so on. I'll give this a go to see if it works. EvanArticle: 20284
We're doing what you want to do. In one project there are 23 libraries and up to 15 developers (including test bench and test code developers). We are using NT and Solaris platforms and sharing the file systems containing the controlled source. The way we did it was to have a template renoirPrefs which the designers copy to their local ~/.renoirPrefs (or the equivalent place for NT). renoirPrefs mostly controls generate styles, editor selection and other stuff users have their own dogmas about. Mistakes in renoirPrefs generally just prevent developers from proceeding with their task until they figure it out. The project is not messed with. We've never had to change renoirPrefs during the project. The renoir.ini (which contains the lib mappings mostly) is selected by setting the $RENOIRLIBS variable in each users environment to the same value. Changing libraries around or adding them requires the renoir.ini be checked out and changed by hand. This does not happen often, and can usually be managed in a backward compatible way. We're using clearcase for source control, and while Renoir can interwork with it, renoir 99.2 makes messes during check in so we do it manually. We use release labels so that individuals can select tested working versions of everything but the stuff they're directly working on. There's a pre-release mechanism used to regress newly checked in source against the rest of the system before it's released for generaly use. SW folks mastered all this sort of stuff a long time ago and we just do what they do. eml@riverside-machines.com.NOSPAM wrote: > > 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. The common renoir.ini refers to all the libraries, so all developers can see all the libraries all the time, even if they are only working on stuff in one of them. If you want to test against a complete system, you'd need all the libraries available anyway. > > 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. Why not just check out the files you need? You generally don't need the all the source files in a library. To make this work, people will be generating their own local copies of the generated files and vhdl libs. In Clearcase each user can have his own set of private files which have the same path name as another user's, but completely different content. Only files explicitly made into source controlled elements are common to everyone. If your source control can't do this, then the paths for the generated and vhdl libraries in renoir.ini will have to have a user-unique variable. e.g. [renoir_library] lib_1 /source_control_file_system/$PROJECT/lib_1/renoir lib_2 /source_control_file_system/$PROJECT/lib_2/renoir [generated_library] lib_1 $USER/objects/lib_1/vhdl.gen lib_2 $USER/objects/lib_2/vhdl.gen [library] lib_1 $USER/objects/lib_1/vhdl.lib lib_2 $USER/objects/lib_2/vhdl.lib Clearcase can use derived objects to permit sharing of object files, but we never bothered in HW because the build times are not that long, and each user only needs about 1 Gbyte of private storage for objects. PHilArticle: 20285
I believe (if i remember correctly), that counting 1's in an "intense" problem in planar vlsi because the circuit complexity is at least as great as parity (remember the parity function is just the lowest bit of a "bitcount" function), and parity circuit complexity is similar to multiplication. And we have non-trivial lower bounds on circuit complexity for mulutiplication (courtesy Andy Yao). So, just warning you that this one of the more difficult problems in VLSI and don't expect to get a screaming fast solution. A common software algorithm is to build a tree of recursive adders to add every pair of bits (in parallel), then add every pair of sums (in paralle), add every pair of sums ... until you get just one sum. Depth is O( log(bits) ), and the adders can be very stupid. No adder needs to be more than log(bits) in size. A more complex software algorithm is to use subtractors and add groups of 3 bits in the first stage. If x = 4a + 2b + c, then x - x>>1 - x>>2 = a+b+c. Then, you have a sum (0..3) and this is in the right range to minimize gates in all the downstream stages. - DonArticle: 20286
Hi - Those of you in the Bay Area who are interested in bypassing and other power distribution issues may be want to attend next week's IEEE EMC Society meeting. The notice that appears below was posted to the signal integrity reflector this AM. Take care, Bob Perlman -------------------------------------------------------------------------------------- The Santa Clara Valley Chapter of the EMC Society is presenting on the 8th of February, 2000 Simultaneous Switch Noise and Power Plane Bounce, by Larry Smith, Sun Microsystems. Free admission, all welcome. 5:30-7:30 pm. social, food and refreshements (there is a charge for food and refreshements) 7:30-8:30 pm. Presentation: Simultaneous Switch Noise (SSN) has traditionally been thought of as an inductance problem. Modern electronic packages with solder bumps, solder balls and power planes have very low inductance. The SSN problem is shifting from an inductance problem to a power plane bounce problem. Return current from signal transmission lines can be used to explain and account for power plane bounce. Noisy power planes are known to be the root cause of many SI and EMI problems. The key to managing power plane bounce is in managing return currents and power plane decoupling. Larry D Smith received the BSEE degree from Rose Hulman Institute of Technology in 1975 and the MS dregree in Material Science from the University of Vermont in 1983. After joining IBM in 1978, he worked in the areas of reliability, characterization, failure analysis, power supply and analog curcuit design, packaging and signal integrity at Sun Microsystems since 1996. His current area of concentration is design of power distribution systems and reduction of simultaneous switch noise. SCV IEEE EMC Society Dinner & Meeting Locations The dinner and meeting will be held at the Silicon Graphics’ Café Iris, Building 5, on 2025 Stierlin Court in Mountain View, CA. For additional information and directions visit our website: www.scvemc.org ----------------------------------------------------- Bob Perlman Cambrian Design Works Digital Design, Signal Integrity http://www.best.com/~bobperl/cdw.htm Send e-mail replies to best<dot>com, username bobperl -----------------------------------------------------Article: 20287
Just found out that RS have a good range also. Stewart, Nial [HAL02:HH00:EXCH] <stewartn@europem01.nt.com> wrote in message news:3899A1FE.E0939579@europem01.nt.com... Tim Tyler wrote: > > 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 :-| One thing I would add to this is check prices before you decide what you're going to use. I was using EPM7096's which aren't isp compatible and which cost £25 until I found I could buy the bigger EPM7128S which is isp compatible for £11. You can also get them through Farnell for about the same price. Nial.Article: 20288
In VLSI, counting 1's isn't that bad. It is just a tally adder, which in merged tree form is basically a wallace tree with one bit wide input 'columns'. Delay is a log(n) problem. The classic optimal merged tree has 2x2 blocks (2 input gates, each fanning out to 2 destinations). The FPGA structure usually supports more than that, so you can improve it by taking advantage of the architecture. Also, the fast carry chains in the FPGA skew the optimal solution. Donald Gillies wrote: > I believe (if i remember correctly), that counting 1's in an "intense" > problem in planar vlsi because the circuit complexity is at least as > great as parity (remember the parity function is just the lowest bit > of a "bitcount" function), and parity circuit complexity is similar to > multiplication. And we have non-trivial lower bounds on circuit > complexity for mulutiplication (courtesy Andy Yao). > > So, just warning you that this one of the more difficult problems in > VLSI and don't expect to get a screaming fast solution. > > A common software algorithm is to build a tree of recursive adders to > add every pair of bits (in parallel), then add every pair of sums (in > paralle), add every pair of sums ... until you get just one sum. > Depth is O( log(bits) ), and the adders can be very stupid. No adder > needs to be more than log(bits) in size. > > A more complex software algorithm is to use subtractors and add groups > of 3 bits in the first stage. If x = 4a + 2b + c, then x - x>>1 - > x>>2 = a+b+c. > > Then, you have a sum (0..3) and this is in the right range to minimize > gates in all the downstream stages. > > - Don -- -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: 20289
If you are not getting spartanII in the device list in the design manager, you either didn't install it or you didn't install it right. I think installing the spartan II requires a new CD code which you get from the Xilinx website. If you are still using your original CD key code, the other stuff still works, but you won't see the SpartanII parts. Search on the Xilinx website for SpartanII and look for the installation instructions. IIRC, the procedure is spelled out in excruciating detail. You need SP4 for the timings, and you need to go through the install procedure after patching with SP4 to enable the spartanII (for that you also need the new key code). Nicolas Matringe wrote: > 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 FRANCE -- -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: 20290
> The actual act of changing a serial number can't be illegal > unless Microsoft forbids it in its licence agreement, Even if it is in a license agreement, it does not mean it is illegal. Most license agreements claim you can't disassemble the code, yet the courts have upheld the position that is perfectly legal for you to do. You own the disk, you have the right to use the software. If you are only using one copy of it at a time, and have not provided it to anyone else, from what I understand the current laws to be, you are not violating any laws.Article: 20291
On Thu, 03 Feb 2000 15:56:01 -0500, Rickman <spamgoeshere4@yahoo.com> wrote: [snip] >So I would not recommend that you use multiple large capacitors. Rather >use one large (and 470 uF is really large) value tantalum cap near the >power connector for each supply voltage. Then use a 0.01 to 0.1 uF cap >as close as possible to each power and ground pin pair. I ended up with >some 90 pieces of 0.1 uF caps on my board giving me a total of about 4 - >5 uF per supply voltage in high freq decoupling. And it was a very small >board, PC/104! [snip] Hey Rick, like what you say, except for the bit about the tantalum cap needing to be near the power connector. It really doesn't matter where it goes on the board. As you say, it's only effective at lower frequencies. Also, its ESR will be much higher than the plane impedance. Regards, Allan.Article: 20292
> My opinion is that Xilinx is practicing some serious overkill without > good justification. Their note about not knowing what the different > designs are doing is real, but there are rules of physics that > electronics follow. > I would agree with your opinon. The app note is really stupid. If you would place all the caps araound the device, you are not able to connect the device, or you loose the bottom of the PCB. We're using Xilinx FPGA for a long time (from XC4xxx up to XCV800) and we don't have had any problems with it. The design of the PCB is more important than the cap's araound it. For EMI-Problems not every time a lot of cap's are really good. A very good solution is the use a wide band decoupling. The way for this is: 1. Use multilayer PCB nboards with seperate power planes 2. Create "island's" for the device 3. calculate with the characteristics of the PCB (Er, distance of layer's, etc.) the correct cap's for a wideband filter. The result is, that only one big cap (220 - 470 uF) can be used for the power input of the board. Then, for the island of the FPGA only a group of up to 6 small cap's (10pF...1uF) are neccessary. The wide band decoupling is developed by Prof. Dirks, the calculation program calls "SILENT". If you want more info, just contact him at EMVProf@aol.com. Andreas HeinerArticle: 20293
Hi friends, I am evaluating Synplify now. They have very good RTL viewer. But one of the components, I have found in there I can't find in XILINX primitives. I speak about PMUX. I think this is a priority MUX. If the E input =1 the output is equal D input (Am I right?). But what will be if two E inputs are one? Can anybody send me the LUT or describe the function of this primitive. Sent via Deja.com http://www.deja.com/ Before you buy.Article: 20294
Ray Andraka wrote <389A359A.2CBBEAA9@ids.net>... >If you are not getting spartanII in the device list in the design manager, >you either didn't install it or you didn't install it right. I think >installing the spartan II requires a new CD code which you get from the >Xilinx website. If you are still using your original CD key code, the other >stuff still works, but you won't see the SpartanII parts. Search on the >Xilinx website for SpartanII and look for the installation instructions. >IIRC, the procedure is spelled out in excruciating detail. You need SP4 for >the timings, and you need to go through the install procedure after patching >with SP4 to enable the spartanII (for that you also need the new key code). > >Nicolas Matringe wrote: > >> 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 FRANCE > >-- >-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/~randraka > > Hi everybody, I have exactly the same problem with Foundation 2.1i sp4 when i try to synthesize or to implement a design on a Spartan II. 2 months a go, i got a new CD Key code from the xilinx website to use Virtex E chips. I have no trouble with making a synthesis / implementation on Virtex E Chips. 1 month a go, i tried to make a synthesis on Spartan II chips, this family was not on the list. I tried to get a new CD key code from the xilinx website to use Spartan II, i got exactly the same code as the Virtex -E one. I made a full unistallation / installation of Foundation, then of SP4 with the new CD code, it still doesn't work. for Spartan II, and it still works for Virtex E. (I use Windows NT). What am i doing wrong ? J-P GOGLIO GETRIS S.A. 13 Chemin des Prés 38240 Meylan Tel : (33) 4 76 18 52 10 Fax : (33) 4 76 18 52 01 E-mail : goglio@getris.comArticle: 20295
Ray Andraka wrote <389A359A.2CBBEAA9@ids.net>... >If you are not getting spartanII in the device list in the design manager, >you either didn't install it or you didn't install it right. I think >installing the spartan II requires a new CD code which you get from the >Xilinx website. If you are still using your original CD key code, the other >stuff still works, but you won't see the SpartanII parts. Search on the >Xilinx website for SpartanII and look for the installation instructions. >IIRC, the procedure is spelled out in excruciating detail. You need SP4 for >the timings, and you need to go through the install procedure after patching >with SP4 to enable the spartanII (for that you also need the new key code). > >Nicolas Matringe wrote: > >> 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 FRANCE > >-- >-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/~randraka > > Hi everybody, I have exactly the same problem with Foundation 2.1i sp4 when i try to synthesize or to implement a design on a Spartan II. 2 months a go, i got a new CD Key code from the xilinx website to use Virtex E chips. I have no trouble with making a synthesis / implementation on Virtex E Chips. 1 month a go, i tried to make a synthesis on Spartan II chips, this family was not on the list. I tried to get a new CD key code from the xilinx website to use Spartan II, i got exactly the same code as the Virtex -E one. I made a full unistallation / installation of Foundation, then of SP4 with the new CD code, it still doesn't work. for Spartan II, and it still works for Virtex E. (I use Windows NT). What am i doing wrong ? J-P GOGLIO GETRIS S.A. 13 Chemin des Prés 38240 Meylan Tel : (33) 4 76 18 52 10 Fax : (33) 4 76 18 52 01 E-mail : goglio@getris.comArticle: 20296
Ray Andraka wrote <389A359A.2CBBEAA9@ids.net>... >If you are not getting spartanII in the device list in the design manager, >you either didn't install it or you didn't install it right. I think >installing the spartan II requires a new CD code which you get from the >Xilinx website. If you are still using your original CD key code, the other >stuff still works, but you won't see the SpartanII parts. Search on the >Xilinx website for SpartanII and look for the installation instructions. >IIRC, the procedure is spelled out in excruciating detail. You need SP4 for >the timings, and you need to go through the install procedure after patching >with SP4 to enable the spartanII (for that you also need the new key code). > >Nicolas Matringe wrote: > >> 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 FRANCE > >-- >-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/~randraka > > Hi everybody, I have exactly the same problem with Foundation 2.1i sp4 when i try to synthesize or to implement a design on a Spartan II. 2 months a go, i got a new CD Key code from the xilinx website to use Virtex E chips. I have no trouble with making a synthesis / implementation on Virtex E Chips. 1 month a go, i tried to make a synthesis on Spartan II chips, this family was not on the list. I tried to get a new CD key code from the xilinx website to use Spartan II, i got exactly the same code as the Virtex -E one. I made a full unistallation / installation of Foundation, then of SP4 with the new CD code, it still doesn't work. for Spartan II, and it still works for Virtex E. (I use Windows NT). What am i doing wrong ? J-P GOGLIO GETRIS S.A. 13 Chemin des Prés 38240 Meylan Tel : (33) 4 76 18 52 10 Fax : (33) 4 76 18 52 01 E-mail : goglio@getris.comArticle: 20297
5 A42MX16FPL84 ACTEL 5 A42MX24FPL84 ACTELArticle: 20298
Keith, Get in touch with either MicroCall(now insight) or Avnet in the UK. F2.1i costs just $95 (DS-FND-BAS-PC) and the Vhdl/Vlog at $495 (DS-FND-BSX-PC) Hope this helps, Dave Hawke Xilinx UK Keith Wootten wrote in message ... >Hi > >I've been using Foundation F1.4 for a while, using XC5215 and Spartan >XCS40. I want to change to the 3.3V XCS40XL part, but my software won't >support this part. > >Apparently, there is no upgrade path for F1.4 and I'll have to *buy* >some new software. To cope with both the XC5215 and the XCS40XL parts, >I'll need to spend over GBP1000 - yes, one kilopound. I already spent >over GBP2000 for the F1.4 stuff, and I'm not a volume user. > >Why do they do this? Surely the small company user has *some* value? >The support from the dealer was poor and the promised training sessions >never materialised once the money was paid, so cost of support is no >justification. > >Can anyone recommend a UK dealer who is not a shark? > >Cheers >-- >Keith WoottenArticle: 20299
Thanks Phil - this is very useful. I've got a couple of other questions, if you don't mind: >The way we did it was to have a template renoirPrefs >which the designers copy to their local ~/.renoirPrefs >(or the equivalent place for NT). renoirPrefs mostly >controls generate styles, editor selection and other stuff >users have their own dogmas about. Mistakes in renoirPrefs >generally just prevent developers from proceeding with >their task until they figure it out. The project is not >messed with. We've never had to change renoirPrefs during >the project. Some users want Renoir to control the downstream tools, so they'll be putting simulator and synthesiser options (since each user also generates a local EDIF) into their .renoirPrefs. As far as I can see, this is going to make regression testing and building difficult, since the master scripts must refer to everyone's individual .renoirPrefs. It sounds like you don't have this problem, if individual .renoirPrefs aren't source-controlled - is this right? My inclination is not to allow Renoir to control any downstream tools. I can write scripts that run Renoir in batch mode just to regenerate any VHDL, but I can't see how user X, who wants Renoir to run his flows, can fit into an overall automated project test or build. >The renoir.ini (which contains the lib mappings mostly) >is selected by setting the $RENOIRLIBS variable in each users >environment to the same value. Changing libraries around or adding them requires >the renoir.ini be checked out and changed by hand. This does >not happen often, and can usually be managed in a backward >compatible way. > <snipped> >e.g. > >[renoir_library] >lib_1 /source_control_file_system/$PROJECT/lib_1/renoir >lib_2 /source_control_file_system/$PROJECT/lib_2/renoir > >[generated_library] >lib_1 $USER/objects/lib_1/vhdl.gen >lib_2 $USER/objects/lib_2/vhdl.gen > >[library] >lib_1 $USER/objects/lib_1/vhdl.lib >lib_2 $USER/objects/lib_2/vhdl.lib It sounds like: (a) you're keeping the working source libraries on the server, rather than checking them out into a local directory, and (b) you're not keeping the generated libraries under source control, but simply regenerating as required - is this right? Does $RENOIRLIBS then point to a renoir.ini on the server? It sounds like Clearcase is probably better than MKS. There's no pre-release/release mechanism; I think I'm going to have to bodge this by pre-releasing into an archive branch, testing, and then merging back onto the main trunk. Thanks Evan
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