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 Boescke wrote: > > Does Xilinx plan to use the single transistor SRAM? See <www.mosysinc.com> for > more info about it. > > About about checking out the page byself ? > > The mosys 1T sram is dram with a clever refresh scheme. Nothing that could > be used in a FPGA. As a counterexample, will it be technologically possible to have reconfigurable FPGA-like control system in DRAMs, so that we can get rid of RDRAM, SDRAM, EDO, ZBT etc.? :) UtkuArticle: 29751
HI, Someone answer to my question but i don't understand exactly, i've an error after the beginning of the implementation, the software say to me, input or output PDA missing, or miss as'save', what is PDA or how i can repair this error, because all pins are connected !! SEB -- **************************** Seb@stien.bizland.com ****************************Article: 29752
"A. dhermies" schrieb: > > i used the x-checker with xc9572PC44 > and ..it doesn't work well. > JTAG is not good! (The IDcode is wrong!) > But the xchecker was good with a virtex... > > I used my home made // cable, and it was the same... Even with a new chip! > > Can the PC be the cause of this problem? > > I don't think I have done mistakes on the board, and electric signals are visible on logic analyser... > Can someboy help me? > thank you. Once i faced problems with different LPT Modes, try a more conservative one regards Bertram -- Bertram Geiger, bgeiger@aon.at HTL Bulme Graz-Goesting - AUSTRIAArticle: 29753
Hello Sir, What you probably mean is that the Input/Output PADs are missing, which would result in your entire design getting trimmed out during map. There's a switch in ngdbuild (-a) that inserts IO pads on all top-level ports. Ensure that the switch is turned on if your synthesis tool does not do this insertion for you. The switch can also be set in the GUI in the translate section of the implementation options. Hope this helps. Regards, Tim Seb C wrote: > HI, > > Someone answer to my question but i don't understand exactly, i've an error > after the beginning of the implementation, the software say to me, input or > output PDA missing, or miss as'save', what is PDA or how i can repair this > error, because all pins are connected !! > > SEB > > -- > > **************************** > Seb@stien.bizland.com > ****************************Article: 29754
Utku, The DRAM process is different from the CMOS process, and making an IC with logic and DRAM is a significant challenge (foundries and companies have failed trying to do so). One can not take advantage of the latest technology, and one finds oneself a few years behind (i.e. death of company to be late!!!) Moore's Law goes: 0.18 -> 0.15u -> 0.12u -> 0.10u -> 0.07u -> ??? If you are not making silicon at the tiny end by now....it may be too late for you as a chip vendor. When it is a standard offered process (DRAM + logic), or even a prototype process that is available from a foundry in the world, we will be ready, Austin Utku Ozcan wrote: > Tim Boescke wrote: > > > > Does Xilinx plan to use the single transistor SRAM? See <www.mosysinc.com> for > > more info about it. > > > > About about checking out the page byself ? > > > > The mosys 1T sram is dram with a clever refresh scheme. Nothing that could > > be used in a FPGA. > > As a counterexample, will it be technologically possible to have > reconfigurable FPGA-like control system in DRAMs, so that > we can get rid of RDRAM, SDRAM, EDO, ZBT etc.? :) > > UtkuArticle: 29755
Eric Smith <eric-no-spam-for-me@brouhaha.com> writes: > Is there any document that gives more detailed logic drawings of > the Spartan II CLBs? Get the Virtex data sheet (Virtex are bitstream-identical with Spartan-II, for equivalent sizes, so the CLB is identical): http://www.xilinx.com/partinfo/ds003.pdf Page 7, Figure 5 is what you want to look at. > The description and diagram in the data sheet > are sort of, well, spartan. :-) > I'm trying to figure out whether I can do certain things like > simultaneously route the carry out both to the carry-in of the CLB above > *and* to the GRM or to the CLB on the right. Yes, both of these connections are fixed wired actually. > These details aren't considered proprietary trade secrets, are they? No. They are published. -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng FH/BSc, Sysadmin, Roleplayer, LARPer, MysticArticle: 29756
> Laurent for Amontec > www.amontec.com > ----------------------------------- > Amontec introduces a new easy PCI development system. > Build your own PCI board in 2 weeks! A reality with Amontec. I went to that web site, and there was really no real information on this 'product'. The products section said 'We are working on this page.", and the front page, though it mentioned this PCI product, has no links that I could find for any more information. Perhaps I am looking in the wrong place, would you mind posting more information on this 'product'?Article: 29757
Chris Dunlap <cdunlap@xilinx.com> writes: > You can always look in FPGA editor. Nothing can be left out there. If its > routed or routable, its there. I don't seem to have an FPGA editor. Is it supposed to be available in WebPACK ISE, or is it only in the "real" software? Thanks! EricArticle: 29758
The CHES registration is up and running. You can go to the CHES main page at www.chesworkshop.org and download the registration form. A list with hotels with special CHES rates will become available very soon. Regards, Christof PS: For those who want to make travel plans: The CHES 2001 program starts on the morning of Monday, May 14, with pre-registration on Sunday evening. CHES ends on Wednesday, May 16, at NOON. ! WORKSHOP ON CRYPTOGRAPHIC HARDWARE AND EMBEDDED SYSTEMS (CHES 2001) ! ! Paris, France, May 13-16, 2001 ! ! www.chesworkshop.org ! *********************************************************************** Christof Paar, Assistant Professor Cryptography and Information Security (CRIS) Group ECE Dept., WPI, 100 Institute Rd., Worcester, MA 01609, USA fon:(508) 831 5061 email: christof@ece.wpi.edu fax:(508) 831 5491 www: http://www.ece.wpi.edu/People/faculty/cxp.html ***********************************************************************Article: 29759
While the approach that's described in the some of these other posts is sound, I've found it's easier and more convenient to form a boundary between the PC/EPP port farther into the logic than has been suggested. The port decoding can be combinatorially done, with the exception of the handshake timing which requires only two flip-flops in order to generate a valid RDY signal. If you use transparent latches for the data and addresses in your primary interface, the downstream registers (edge-triggered with the decoded port signals) will always have "at least plenty" of setup and hold time because you use the rising edge of RDY to clock them and data is always valid before the occurrence of the first strobe (nDS, nAS, nWS). That reduces the logic, ensures you have sufficient setup, and since you're latched, you have essentially infinite hold time. The EPP standard that's been floating around for some time is paced to match the ISA transfer rates, so timing is exceedingly generous. There's no telling what the behavior of the true PCI EPP will be, though and it may require a mite more care. Dick On Mon, 05 Mar 2001 10:11:55 +0100, Marc Reinert <reinert@tu-harburg.de> wrote: >I' like to use a CPLD/FPGA (Xilinx) to receive data from the parallel >port (EPP-mode) of my PC. > >Is it a good style to react direct on the edges of the port signals (e. >g. adress/data strobes) or would it be better to use a fast PLD-Clock to >sample the port and then to evaluate the signals in a clocked logic? > >Marc > >Article: 29760
We designed a product around the Spartan II PWDN (powerdown) pin and are in production. I just downloaded a new version of the data sheet and found that Xilinx deleted all reference to the PWDN pin on 11/02/00 from all the datasheets. We are screwed if this is true. Anyone have any info? There is nothing in the answers database.Article: 29761
Dan Rudolf wrote: > We designed a product around the Spartan II PWDN (powerdown) pin and are in production. I just downloaded a new version of the data sheet and found that Xilinx deleted all reference to the PWDN pin on 11/02/00 from all the datasheets. We are screwed if this is true. Anyone have any info? There is nothing in the answers database. My distributor told me a few months ago that the power-down pins were being removed from Spartan2 and the modification and production ramp-up time for the modified chip was the reason for their shortage. It's anecdotal, but it correlates with your experience. I wasn't monitoring the Xilinx announcements on Spartan2 but I assume they made it known well in advance that this change was coming, although I expect you have a very different opinion on that. -- || 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 ||Article: 29762
You're screwed. The powerdown is only in the ES (engineering sample) parts. I thought I had seen something in the answers database about a year ago on it. Can't find it right now though. The database page is not responding. IIRC, there was a yield issue. Perhaps it is related to the delay in availability of parts. Dan Rudolf wrote: > > We designed a product around the Spartan II PWDN (powerdown) pin and are in production. I just downloaded a new version of the data sheet and found that Xilinx deleted all reference to the PWDN pin on 11/02/00 from all the datasheets. We are screwed if this is true. Anyone have any info? There is nothing in the answers database. -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com or http://www.fpga-guru.comArticle: 29763
On Wed, 07 Mar 2001 12:44:45 -0800, Austin Lesea <austin.lesea@xilinx.com> wrote: >Utku, > >The DRAM process is different from the CMOS process, and making an IC with logic and >DRAM is a significant challenge (foundries and companies have failed trying to do >so). One can not take advantage of the latest technology, and one finds oneself a few >years behind (i.e. death of company to be late!!!) > >Moore's Law goes: 0.18 -> 0.15u -> 0.12u -> 0.10u -> 0.07u -> ??? > >If you are not making silicon at the tiny end by now....it may be too late for you as >a chip vendor. > >When it is a standard offered process (DRAM + logic), or even a prototype process that >is available from a foundry in the world, we will be ready, Well... http://www.chips.ibm.com/products/asics/products/edram/index.html Bring cash. ;-) ---- Keith R. Williams IBM Microelectronics (however, not speaking for IBM - don't do ASICs)Article: 29764
Eric Smith wrote: >Chris Dunlap <cdunlap@xilinx.com> writes: >> You can always look in FPGA editor. Nothing can be left out there. If its >> routed or routable, its there. > >I don't seem to have an FPGA editor. Is it supposed to be available >in WebPACK ISE, or is it only in the "real" software? > WebPACK 3.2WP3.0 includes neither the FPGA Editor nor the Xilinx Design Language (XDL) ASCII <-> NCD conversion utility. Other than not being able to see what the CLB's look like, these omissions also make it very difficult to find and work around the inevitable tool bugs. ( But it's free! ) > Is there any document that gives more detailed logic drawings of > the Spartan II CLBs? Other than the slice drawing in the Virtex datasheet, slightly more information can also be found in the 3.x Libraries Guide "Design Elements" chapters, under MULT_AND, XORCY, and MUXCY. ( Unfortunately, the section titled "Carry Logic in Other Architectures" in the "Attributes, Constraints, and Carry Logic" chapter is also quite spartan, consisting of only a paragraph or so for Spartan-II). Brian DavisArticle: 29765
WE ARE SCREWED!!!!! I found a solution record: http://support.xilinx.com/xlnx/xil_ans_display.jsp?iLanguageID=1&iCountryID=1&getPagePath=10500 We have been using the production XC2S150's with the power down since September. I am REALLY pissed at Xilinx because we went through a large amount of back and forth on exactly how much power the chip would use during powerdown because it was very important in our project. We finally got them to commit on a number and then looking at the real parts it was about 10X less. We were fat dumb and happy..... until today. Dan Dan Rudolf Sapphire Computers, Inc. (937) 767-1062Article: 29766
Hi, Is there anyone I can talk to about FPGA development for GSM baseband chipsets?? please respond to chris@goldgroup.co.uk Regards ChrisArticle: 29767
Keith, Thanks. I sat in on the presentation from IBM to Xilinx. The 0.18u is already two years old, but the .11u does look interesting. http://www.xilinx.com/prs_rls/ibmpartner.htm I suppose if you actually read all the press releases, and read between the lines, one could almost guess at where things might be going. It is pretty wild, that is for sure. Austin "Keith R. Williams" wrote: > On Wed, 07 Mar 2001 12:44:45 -0800, Austin Lesea > <austin.lesea@xilinx.com> wrote: > > >Utku, > > > >The DRAM process is different from the CMOS process, and making an IC with logic and > >DRAM is a significant challenge (foundries and companies have failed trying to do > >so). One can not take advantage of the latest technology, and one finds oneself a few > >years behind (i.e. death of company to be late!!!) > > > >Moore's Law goes: 0.18 -> 0.15u -> 0.12u -> 0.10u -> 0.07u -> ??? > > > >If you are not making silicon at the tiny end by now....it may be too late for you as > >a chip vendor. > > > >When it is a standard offered process (DRAM + logic), or even a prototype process that > >is available from a foundry in the world, we will be ready, > > Well... > > http://www.chips.ibm.com/products/asics/products/edram/index.html > > Bring cash. ;-) > > ---- > Keith R. Williams > IBM Microelectronics > (however, not speaking for IBM - don't do ASICs)Article: 29768
Hello everybody, I just installed Xilinx 3.3i Service Pack 7 for Alliance-PC, and it crashes every time I use the mapper on Virtex-E parts. Seems to be OK on Spartan. I get the good ol' Windows error box and a message: "protection fault in module LIBLMCIRCUIT.DLL". Doesn't matter if I execute the mapper from command prompt or the flow engine. It stops at exactly the same point, after "Reading NGD file...". Funny thing is, my colleague installed the exact same upgrade on his (identical) PC and it works just fine. I've done (several!) clean installs from the original CDs, upgrading through 3.1, 3.2sp5, 3.3sp6 etc and every version works perfectly OK until service pack bloody 7 which crashes consistently in the MAP. Anybody any ideas? Have I got the wrong version of MFC42.DLL or some such thing???? Thanks Mike H.Article: 29769
I made simple design: put two COMPCM8 macros, attached RLOC attributes to them (let's say: R0C0.S0 and R10C0.S0), then I created a Macro Symbol from this schematic, put the macro onto top schematic, attached an RLOC_ORIGIN attribute (R10C10), and run the implementation. During MAP I got 32 errors like: ERROR:Map:1 - "RLOC" suffix on MUXCY_L symbol "H1/$I51/$1I1710" (output signal=H1/$I51/$1I1710/O) conflicts with one higher in the design. and a final error: ERROR:Map:52 - Problem encountered processing RPMs. What am I doing wrong with RLOC's ? Regards, Pawel J. RajdaArticle: 29770
Has anyone managed to get (or at least order) a Foundation ISE Evaluation Kit? According to a Xilinx press release of 21-Feb-2001, it is available free for a 60-day trial period. But there's no part number given, and my distributor hasn't heard of it. Thanks! EricArticle: 29771
I wrote: > Has anyone managed to get (or at least order) a Foundation ISE Evaluation > Kit? According to a Xilinx press release of 21-Feb-2001, it is available I should have included the URL: http://www.xilinx.com/prs_rls/0111iseeval.htmArticle: 29772
Hi all target Xilinx FPGA. If a signal is routed through the longline, does the high fanout of this signal, affect its final speed, or no ? Say a transpose filter structure , the input is broadcasted to all the multpliers, i know this architecture is more optimised, at least in term of area, than the direct form, but will this long line broadcast not affect the final performance ? If we have a high fanout signal coming from the CLB output. How far the degradation will be in function of of the fanout. can this output be swicthed to the longline ? <~> Peter <~> __________________________________________________ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/ -- Posted from web1602.mail.yahoo.com [128.11.23.202] via Mailgate.ORG Server - http://www.Mailgate.ORGArticle: 29773
Have you tried more than one design? Could be something in your design the mapper doesn't like. "Mike H." wrote: > > Hello everybody, > > I just installed Xilinx 3.3i Service Pack 7 for Alliance-PC, > and it crashes every time I use the mapper on Virtex-E parts. > Seems to be OK on Spartan. > > I get the good ol' Windows error box and a message: > "protection fault in module LIBLMCIRCUIT.DLL". > > Doesn't matter if I execute the mapper from command > prompt or the flow engine. It stops at exactly the same > point, after "Reading NGD file...". > > Funny thing is, my colleague installed the exact same > upgrade on his (identical) PC and it works just fine. > > I've done (several!) clean installs from the original > CDs, upgrading through 3.1, 3.2sp5, 3.3sp6 etc > and every version works perfectly OK until > service pack bloody 7 which crashes consistently > in the MAP. > > Anybody any ideas? Have I got the wrong version of > MFC42.DLL or some such thing???? > > Thanks > > Mike H. -- -Ray Andraka, P.E. President, the Andraka Consulting Group, Inc. 401/884-7930 Fax 401/884-7950 email ray@andraka.com http://www.andraka.com or http://www.fpga-guru.comArticle: 29774
Brian, Try isolating INIT and DONE pins of Spartan-II and Spartan-XL devices. This helps in debugging daisy chain configurations and we can figure out which device is failing configuration by monitoring INIT pins separately. It seems like you are able to configure the Spartan-XL device but not Spartan-II. Spartan-II devices configuration starts after it sees the 32-bit sync word (0xAA995566). Probe the DOUT of Spartan-XL to see if its passing this sync word onto Spartan-II device. You may also try using a 4.7Kohm resistor on INIT and 300 ohm resistor on DONE pin of the Spartan-II device. Another good resource would be to check out Configuration problem solver on Xilinx support website at http://support.xilinx.com/support/troubleshoot/psolvers.htm Hope this helps !!! -Vikram Brian Carruthersq wrote: > I have an application that uses a 95144XL as an interface from firmware to program a XCS40XL and a XC2S150 in Slave serial mode. > > The Spartan XL is the lead device in the daisy chain to the Spartan II. The bitstreams for the two devices are combined using the prom generator. After the firmware has sent all of the configuration data the DONE line stays low, and the INIT stays high. > > To verify that the logic in the CPLD and the functionality of the firmware, I removed the Spartan II from my PCB. We re-ran the program to only send the bitstream for the Spartan XL. This works perfectly. > > The problem seems to be that when the Spartan II is on the PCB it keeps the DONE low. > > The configuration for the Spartan II is set so that the DONE has the Active pullup option selected. > > Any suggestions that you may have would be appreciated.
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