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
I am interested in finding and/or setting-up a GNU-style library of free synthesisable IP cores. It seems to me that the GNU GPL licensing model (free source, do what you want with it but you must acknowledge it and keep it complete) has some potential benefits particularly for small businesses using FPGAs. One good aspect of GNU GPL is that it doesn't prevent people from making money out of consultancy and support activities around the free material. Another is that it might make available some cores that are too small to be of interest to serious IP companies. And yes, I do know about ACTgen and Xilinx's core generator, which are fine, but don't cover everyone's needs. If such a thing exists already, can anyone point me at it please? And if it doesn't, I would be willing to put some work in to setting it up and hosting it - anyone interested? I even have an acronym already: ELK (Electronics Library of Kernels) - it is of course inappropriate to have a recursive acronym like GNU because all hardware designs must be statically elaborated :-) Jonathan BromleyArticle: 16801
Try http://www.cmosexod.com/ http://circu.its.tudelft.nl/ http://www.scsise.wmin.ac.uk/~seamang/freehardware.html http://www.scrap.de/html/openip.htm http://www.geocities.com/SiliconValley/Chip/5014/ http://www.vhdl.org/vi/fmf/ http://tech-www.informatik.uni-hamburg.de/vhdl/vhdl.html Post any more you find... Have Fun Stephen Smith Jonathan Bromley wrote: > > I am interested in finding and/or setting-up a GNU-style library > of free synthesisable IP cores. It seems to me that the GNU GPL > licensing model (free source, do what you want with it but you > must acknowledge it and keep it complete) has some potential benefits > particularly for small businesses using FPGAs. One good aspect > of GNU GPL is that it doesn't prevent people from making money > out of consultancy and support activities around the free material. > Another is that it might make available some cores that are too small > to be of interest to serious IP companies. > > And yes, I do know about ACTgen and Xilinx's core generator, which > are fine, but don't cover everyone's needs. > > If such a thing exists already, can anyone point me at it please? > And if it doesn't, I would be willing to put some work in to > setting it up and hosting it - anyone interested? > > I even have an acronym already: ELK (Electronics Library of > Kernels) - it is of course inappropriate to have a recursive > acronym like GNU because all hardware designs must be statically > elaborated :-) > > Jonathan BromleyArticle: 16802
That works too, and you can use common rw controls if you write two words in one cycle and read two in the next. It costs a pair of alignment registers in the FPGA, but frees up a few pins. Lasse Langwadt Christensen wrote: > how about two ram chips, connect the addresses but "swap" read and write > on the two chips, then you can read from chip 1 at the same time as you > write > to chip 2, once you've got to the length of the buffer you start reading > from 2 > and writing to 1 ? you'll need twice as many data io pins and twice as > much mem > but it'll only have to run at single speed > > --L2C > --___--_-_-_-____--_-_--__---_-_--__---_-_-_-__--_---- > Lasse Langwadt Christensen, MSEE (to be, june 30th 1999) > Aalborg University, Department of communication tech. > Applied Signal Processing and Implementation (ASPI) > http://www.kom.auc.dk/~fuz , mailto:langwadt@ieee.org -- -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: 16803
Has anyone come across any guidelines for how many simultaneous switching outputs the Spartan XL's can tolerate ? They published an app note about the Virtex, but there does not appear to be anything on the Spartans. Some of Xilinx's CPLD parts had a "programmable ground" feature that drives unused I/Os low to act as pseudo-ground pins. Anyone ever try this on an FPGA to see if it reduces switching noise ? -- ----- Terry Fraser Hardware Designer Applied Microelectronics Inc Halifax, NS Phone: (902) 421-1250 ext 269 FAX: (902) 429-9983 mailto:fraser@appliedmicro.ns.ca http://www.appliedmicro.ns.caArticle: 16804
The Qualis special workshop on Design Reuse is only a week away! With attendance nearly full, you have only a few more days to reserve your seat at this enlightening event. The details follow, but call or register online soon! Designing with Reuse in Mind A Special 3-day Workshop June 16-17-18 1999 San Jose, California Qualis Design is hosting a special workshop on design reuse this June 16-17-18 in San Jose, California. This intense workshop will be led by Janick Bergeron, Vice President of Technical Wisdom at Qualis. Janick will work closely with the attendees of this stimulating workshop, covering all of the critical areas of design reuse. As a recognized expert in reuse methodologies and large system design, Janick deeply understands the reuse issues designers face. In this interactive workshop, you will learn how to set up a reuse-friendly design environment, how to architect, implement, and verify a reusable component and SoC design, and how to package it for future reuse. Join other top engineers for this unique opportunity to network and discuss today's reuse challenges and solutions. Find detailed information about this seminar on our website at: http://www.qualis.com/cgi-bin/qualis/session.pl?event=reuse_990616.html or call Michael McDowell at 888-214-4224 (or +1-503-968-8550) to secure your seat. MichaelArticle: 16805
Hal Murray <murray@pa.dec.com> wrote (quoting me): :> Say I am implementing a lattice-gas automata in order to simulate :> turbulent fluid flow through a confined space. What earthy use are :> counters or addition and subtraction primitives to me? None whatsoever. :> :> If I want to retain the flexibility to run more than one type of :> lattice-gas simulation, FPGAs are the best type of target hardware. : How much money to you have? As always, not /quite/ as much as I'd like ;-) : Have you considered making an ASIC that was a big array of FPGA : type cells optomized for your problem? Yes, though I don't feel able to draw any firm conclusions at this stage. In fact I have several "target problems": one involves a very simple automata capable of universal computation (in fact /almost/ identical to Margolus's implementation of Fredkin's "billiard-ball machine": http://www.alife.co.uk/ca/bbm/2d/) - and the other is an unusual and complex automata which supports self-reproduction: http://www.alife.co.uk/hal/ This model is not yet at all fixed; and may require ongoing modifications. It seems to me that if I *can* afford lots of silicon, then prototyping a circuit on an FPGA, allowing my model to settle down, and then making my own chips consisting of hardware devoted to running my automata - may well wind up being cheaper. If I *can't* afford lots of silicon, then I'm best off getting the biggest and bestest FPGA that I can afford; as it's all I'm going to get. -- __________ |im |yler The Mandala Centre http://www.mandala.co.uk/ tt@cryogen.com When I gave her the ring, she gave me the finger.Article: 16806
Peter Alfke wrote > > The pull-up "resistor", which of course is a p-channel transistor, is active > even for voltages above Vcc. So you should count on a "resistor" to the Vcc > terminal of about 50 kilohm. It's not "leakage", but rather a non-linear > resistance. > I have been corrected by the circuit designer: Once the pin voltage is one threshold above Vcc, the pull-up "resistor" is actually turned off, so the current disppears (down to the microamp or below level). Trying to be honest. Peter Alfke, Xilinx ApplicationsArticle: 16807
NEC and others make video line delay RAMs. If (1-3 line) video line delay is what you want, this is an inexpensive and off-the-shelf solution. Inherent in this approach is that there is no random access; once you pop the next data word out of the FIFO, the previous word is gone (unless you've cached it on the FPGA in a small local pixel buffer). -- Bob Elkind khalid wrote: > > Hi All, > > I want to implement a line delay using off-chip RAM (necessary for Image > Convolution). On the FPGA, this is done using Synchronous RAM which has a > input data line and an output data one, which makes things easy. The line > delay I want to implement is big (800 pixels) and cannot be implemented on > the FPGA chip. How could I use the off-chip RAM to do that (One line for > both input and output) ?. > > I hope I'll have a reply soon. > > Thanks in advance.Article: 16808
NEC and others make video line delay RAMs. If (1-3 line) video line delay is what you want, this is an inexpensive and off-the-shelf solution. Inherent in this approach is that there is no random access; once you pop the next data word out of the FIFO, the previous word is gone (unless you've cached it on the FPGA in a small local pixel buffer). -- Bob Elkind khalid wrote: > > Hi All, > > I want to implement a line delay using off-chip RAM (necessary for Image > Convolution). On the FPGA, this is done using Synchronous RAM which has a > input data line and an output data one, which makes things easy. The line > delay I want to implement is big (800 pixels) and cannot be implemented on > the FPGA chip. How could I use the off-chip RAM to do that (One line for > both input and output) ?. > > I hope I'll have a reply soon. > > Thanks in advance.Article: 16809
Dear all, sorry for the off-topic posting but we are desperately looking for some Xilinx XCV300 in the PQ240 package. The faster the better, but any speed grade will do. The current lead time from the distributors is more than 10 weeks, so we would be very glad if anyone could sell us some of these parts. Thanks, -Arrigo -- Dr. Arrigo Benedetti o e-mail: arrigo@vision.caltech.edu Caltech, MS 136-93 < > phone: (626) 395-3695 Pasadena, CA 91125 / \ fax: (626) 795-8649Article: 16810
For Nasty Sex links visit: http://website.lineone.net/~simonsi/sexlinks.htm cgxrpvdjtxiuryjktzmrocvbkylngclxfrfbylflmxxzbuuiqutqgArticle: 16811
For Nasty Sex links visit: http://website.lineone.net/~simonsi/sexlinks.htm lrpvmcioypvohcoggskypdwlmtlwhgqssfboekfegttbpcuxbtwrwsqrmuftzephpelstgtctinzpgupbnmhxjArticle: 16812
Peter Alfke wrote: > > I have been corrected by the circuit designer: > > Once the pin voltage is one threshold above Vcc, the pull-up "resistor" is > actually turned off, so the current disppears (down to the microamp or below > level). > For the process technology that the Spartan series is made with, what is Vt for the P channel devices? -- Brian C. Boorman Harris RF Communications Rochester, NY 14610 XYZ.bboorman@harris.com <Remove the XYZ. for valid address>Article: 16813
Hi folks, does anyone out there have information how to configure io-pads with maxII. I'm wondering how to enable/disable output-ff - for example. Is there any information on the web about the available settings in maxII - the online help doesn't really help on this. Much thanks TomArticle: 16814
Brian Boorman wrote: > Peter Alfke wrote: > > > Once the pin voltage is one threshold above Vcc, the pull-up "resistor" is > > actually turned off, so the current disppears (down to the microamp or below > > level). > > > > For the process technology that the Spartan series is made with, what is > Vt for the P channel devices? > Vt is ~700 mV Peter AlfkeArticle: 16815
I'm looking at buying a Virtex board (preferably with a PCI interface) and I'd be interested to here what people think of the Virtex boards that they have used. Cheers -- --------------------------------------------------------------------------------------------- David Braendler http://gene.bsee.swin.edu.au/daveb/index.htm Centre for Intelligent Systems Swinburne University of Technology --------------------------------------------------------------------------------------------Article: 16816
test kfyfeshvoxyhcxdkequmdovdcgfuqtjoyjzqeqwystfepwbobywvzxtgtqkbuuoyrgzkpldftjkkjhtoqieArticle: 16817
On Wed, 09 Jun 1999 14:06:29 -0700, Peter Alfke <peter@xilinx.com> wrote: >> The pull-up "resistor", which of course is a p-channel transistor, is active >> even for voltages above Vcc. So you should count on a "resistor" to the Vcc >> terminal of about 50 kilohm. It's not "leakage", but rather a non-linear >> resistance. > >I have been corrected by the circuit designer: > >Once the pin voltage is one threshold above Vcc, the pull-up "resistor" is >actually turned off, so the current disppears (down to the microamp or below >level). Thanks Peter. Do you mean that the circuit actually has a negative resistance region? Does the following curve represent what really happens? If so, do you know the maximum current into the device? ^I | | | /| | / | +--------+--+--+------>V | /3 4 5 | / | / | / |- | (The weak "hold" circuits on 74LVC logic also exhibit this effect, see http://www-us.semiconductors.philips.com/acrobat/applicationnotes/AN2022.pdf (figures 4 and 5, ignoring the pull-down component) for an example. BTW, I assume that this shape also applies to the strong p channel pull-ups. Regards, Allan.Article: 16818
1. Select the I/O pin. 2. Assign.Logic Options.Individual Logic Options.Fast I/O (on/off) Hope this helps Markus Thomas Rathgen wrote: > Hi folks, > > does anyone out there have information how to configure io-pads with > maxII. I'm wondering how to enable/disable output-ff - for example. > Is there any information on the web about the available settings in > maxII - the online help doesn't really help on this. > > Much thanks > > TomArticle: 16819
Hi, I'm trying to use the student edition of Xilinx F1.5. When I place and route, if I've locked all the pin in the design to specific locations the flow engine hangs when it gets to the constructive placer. If I leave at least one pin un-located there is no problem. Is this a limitation of the student edition or a known problem. Has anyone else come across this? Is there a work around other that not locating all the pins. It is a bit of a pain not being able to specify all locations. Thanks, Darrell Gibson. Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.Article: 16820
gibsond@bournemouth.ac.uk wrote: > I'm trying to use the student edition of Xilinx F1.5. When I place and > route, if I've locked all the pin in the design to specific locations > the flow engine hangs when it gets to the constructive placer. If I > leave at least one pin un-located there is no problem. Is this a > limitation of the student edition or a known problem. Has anyone else > come across this? Is there a work around other that not locating all > the pins. It is a bit of a pain not being able to specify all > locations. Duh. I think I can figure out where that single remaining I/O would be placed, given that only one pin isn't locked. - ReinoudArticle: 16821
www.edy.net/jnoniaArticle: 16822
Reinoud wrote: > > gibsond@bournemouth.ac.uk wrote: > > I'm trying to use the student edition of Xilinx F1.5. When I place and > > route, if I've locked all the pin in the design to specific locations > > the flow engine hangs when it gets to the constructive placer. If I > > leave at least one pin un-located there is no problem. Is this a <snip> > Duh. I think I can figure out where that single remaining I/O would > be placed, given that only one pin isn't locked. > > - Reinoud I think you assume too much. He may not be using all the pins of the device, so leaving it unlocked would allow the placer to put it on whatever unused pin it felt like using. Another work-around is to create a dummy signal pin that does something trivial and let that pin that you don't care about be the "floater". I would try searching the online support section at www.xilinx.com and failing that call the tech support line and pose it to them. Xilinx is one of the best in terms of technical support (IMHO). -- Brian C. Boorman Harris RF Communications Rochester, NY 14610 XYZ.bboorman@harris.com <Remove the XYZ. for valid address>Article: 16823
In article <3760F06A.E0E039FE@casema.net>, Reinoud <dus@casema.net> wrote: > gibsond@bournemouth.ac.uk wrote: > > I'm trying to use the student edition of Xilinx F1.5. When I place and > > route, if I've locked all the pin in the design to specific locations > > the flow engine hangs when it gets to the constructive placer. If I > > leave at least one pin un-located there is no problem. Is this a > > limitation of the student edition or a known problem. Has anyone else > > come across this? Is there a work around other that not locating all > > the pins. It is a bit of a pain not being able to specify all > > locations. > > Duh. I think I can figure out where that single remaining I/O would > be placed, given that only one pin isn't locked. > > - Reinoud > Duh, I'm not using all the pins on the device. If you have a method of guessing the location of pin assignments that tool makes I'd love to hear about it. Darrell Gibson Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.Article: 16824
In article <3760F06A.E0E039FE@casema.net>, Reinoud <dus@casema.net> wrote: > gibsond@bournemouth.ac.uk wrote: > > I'm trying to use the student edition of Xilinx F1.5. When I place and > > route, if I've locked all the pin in the design to specific locations > > the flow engine hangs when it gets to the constructive placer. If I > > leave at least one pin un-located there is no problem. Is this a > > limitation of the student edition or a known problem. Has anyone else > > come across this? Is there a work around other that not locating all > > the pins. It is a bit of a pain not being able to specify all > > locations. > > Duh. I think I can figure out where that single remaining I/O would > be placed, given that only one pin isn't locked. > > - Reinoud > Duh, I'm not using all the pins on the device. If you have a method of guessing the location of pin assignments that tool makes I'd love to hear about it. Darrell Gibson Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.
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