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
On Tue, 20 Jun 2006 06:14:25 +0000, Adam Goldman wrote: > In article <pan.2006.06.11.03.15.36.415791@example.net>, Rich Grise wrote: >>OK, I decided to take a chance and download that 839MB shell script that's >>written for RedHat Enterprise, and was doing OK, (I had to shell out as >>root a couple of times to give the install script permission to write to a >>new directory, but that felt kind of kewl. :-) ), and now I'm at kind of a >>stopper. The graphic install has the progress bar at 99%, and there's a > >>/lib/modules/misc/windrvr6.o: kernel-module version mismatch >> /lib/modules/misc/windrvr6.o was compiled for kernel version >> 2.4.18-14 while this kernel is version 2.4.31. > > Not specifically for your distribution but here's something on recompiling > xpc4drvr and windrvr6 for a different kernel: > > http://gentoo-wiki.com/HOWTO_Xilinx > > It looks to me as if unless you're using their programmer hardware you > don't need to bother. Synthesis and schematic capture ought to work without > this driver. > > Coincidentally, Xilinx mailed out my copy of ISE today, so I'll know > soon enough ;-) > Thanks, but that link only talks about kernel 2.6, but Slackware ships with 2.4, and 2.6 is an option, that I've been afraid to try. ("Compile a new kernel????? Are you nuts?" Thanks! RichArticle: 104251
Rich Grise wrote: > On Tue, 20 Jun 2006 06:14:25 +0000, Adam Goldman wrote: > >> In article <pan.2006.06.11.03.15.36.415791@example.net>, Rich Grise wrote: >>> OK, I decided to take a chance and download that 839MB shell script that's >>> written for RedHat Enterprise, and was doing OK, (I had to shell out as >>> root a couple of times to give the install script permission to write to a >>> new directory, but that felt kind of kewl. :-) ), and now I'm at kind of a >>> stopper. The graphic install has the progress bar at 99%, and there's a >>> /lib/modules/misc/windrvr6.o: kernel-module version mismatch >>> /lib/modules/misc/windrvr6.o was compiled for kernel version >>> 2.4.18-14 while this kernel is version 2.4.31. >> Not specifically for your distribution but here's something on recompiling >> xpc4drvr and windrvr6 for a different kernel: >> >> http://gentoo-wiki.com/HOWTO_Xilinx >> >> It looks to me as if unless you're using their programmer hardware you >> don't need to bother. Synthesis and schematic capture ought to work without >> this driver. >> >> Coincidentally, Xilinx mailed out my copy of ISE today, so I'll know >> soon enough ;-) >> > > Thanks, but that link only talks about kernel 2.6, but Slackware ships > with 2.4, and 2.6 is an option, that I've been afraid to try. ("Compile > a new kernel????? Are you nuts?" > Do you want to use the Xilinx programming tools (impact)? If not, you don't need those two drivers, as Adam said. If you do want to use those tools, then Xilinx requires that you recompile the drivers against your kernel, if you are using a kernel different from the one the compiled against. Kind of annoying, and hopefully one of these days Xilinx will do it right. If you have not recompiled your own kernel, then you probably just need to install the kernel sources corresponding to the kernel you have, and compile the drivers. I don't use Slackware, so I don't know how it is done there. I see that this is crossposted to alt.os.linux.slackware, so someone there could probably help with that.Article: 104252
Steve, That .s/.S proprocessor thing is a feature (.S signals gcc to use the preprocessor, .s tells it not to use the CPP). Assuming all the linkers and such are right (I am not an expert on this by any means), I'd wonder how you are actually compiling and linking. How are you calling the tools? Do you have a makefile? Maybe the .S isn't getting into the action with your makefile? Joey sgfallows@gmail.com wrote: > I'm using a V2Pro based eval board with a simple set of frimware and > software files. The software set includes the Xvectors.S file from the > EDK library and I want/expect the code in the xvectors.S file to be > loaded into memory with the image. > > The vector code is in section .vectors, which is mapped by the linker > script to 0xFFFF0000. However the map file shows zero length for the > section and the debugger confirms that what is loaded there is the next > section after the .vectors section. > > map file snippet: > > .vectors 0xffff0000 0x0 > 0xffff0000 __vectors_start = . > *(.vectors) > 0xffff0000 __vectors_end = . > > linker script snippet: > > .vectors : { > __vectors_start = .; > *(.vectors) > __vectors_end = .; > } > plb_bram_if_cntlr_1 > > > xvectors.S file snippet: > > .section .vectors,"ax" > > .globl _vectorbase > > _vectorbase: > // Vector 0x0000, Jump to zero. > critical_interrupt 0000, 0 > > // Vector 0x0100, Critical interrupt. > critical_interrupt 0100, 1 > > It all looks right to me. I must be missing something. I've tried lots > of name/case/quoting changes. I can't get the code for the vectors > (expanded macro 'critical_interrupt" etc.) to be loaded in memory no > matter what I do. Even messed with the case of the .s file extension > which very bizarrely (to me anyway) changes whether the C preprocessor > is invoked. > > Any clues or ideas?? > Thanks > SteveArticle: 104253
Yes, we support "channel bonding", which means we can use four transceivers, each at 2.5/3.125 Gbps to operate with a throughput of 10 Gbps. But it involve four transceivers and four physical connections. I think Vivek's question was whether the four channels can be combined to one. Physical the answer is "NO". Virtual the answer is "YES", through channel bonding. Peter Alfke ========== Falk Brunner wrote: > Vivek Menon schrieb: > > Hi Sean, > > Can you please elaborate on Virtex-II Pro X and the technique to > > combine 4x2.5 Gbit/sec into 10 Gbit/sec? > > Very simple. The key word is channel bonding. This is done every day in > the 10G XAUI interfaces. So the idea is not new and it is doable without > too much trouble. > > Regards > Falk > > P.S. The 4x2.5G is 4x3.125G due to 8B10B coding.Article: 104254
Hi Andy, as already said by John, it's not done in a single pass. A simple explanation can be found here (regardless of the typo on p.4: read "SZ and SY"): http://web.cecs.pdx.edu/~mperkows/temp/021.Introduction-state-minimization-complete.pdf A more scientific approach can be found here: http://av.rds.yahoo.com/_ylt=A9ibyK.OMZpEXW4AhZSTLaMX;_ylu=X3oDMTBvdmM3bGlxBHBndANhdl93ZWJfcmVzdWx0BHNlYwNzcg--/SIG=12a31utog/EXP=1151042318/**http%3a//web.cecs.pdx.edu/%7ealanmi/510FM2/510OCE-11.ppt While the first paper can be understood and applied even by undergraduates, the second one is very theoretical. Funny coincidence: both explain the rule for equivalent states on p.4 :-) Best regards Eilert Andy schrieb: > I'm not sure I understand how your simplified state table optimization > is going to find identical sequences of states, since the individual > states do not have identical from's and to's compared to any other > individual states, but the sequence as a whole may be identical to > another sequence as a whole. It seems to me that this is where the > most redundant states are going to be found. In software, they usually > write these identical sequences as subroutines, perhaps we should focus > on the same, and use sub-state-machines.Article: 104255
"Ralf Hildebrandt" <Ralf-Hildebrandt@gmx.de> wrote in message news:4ft3joF1joptjU1@individual.net... > Morten Leikvoll wrote: > >> Error from vcom: Attribute "event" requires a static signal prefix. >> Xst handles this very nice.. > > Synthesis with Synopsys DC is even worse. It does not even allow > > my_gen: if N in 0 to 7 generate > ... > if rising_edge(clock_vector(N)) then > ... > end generate; > ..... > You could select the clock-signal out of the vector before calling the > procedure. Then you feed only the selected signal to the procedure. ... > Hmm ... dual-edge behavior? Do you want to model synthesizable code? Have > a look at http://www.ralf-hildebrandt.de/publication/pdf_dff/pde_dff.pdf > if you really need it and want to be independent from the tools. > > > Ralf Looks like an idea.. Can I call a procedure from another procedure? In that case I can try write a new procedure to extract the bit and pass it (as inout?) to a clock detecting procedure using a clean bit. Hmm.. not sure if that will work tho.. Gotta try it. I am not really trying to use dual edge here. I only want to program my cpu_register to be able to detect a rising of falling edge from my system depending on a constant input parameter (here, the input variable 'neg'). Thanks for suggestion and the interesting article :)Article: 104256
Hi Austin, that sounds reasonable. Security proofs are expensive. For the V2 boards you gave away ... what reward did you offer in case of success? I suspect there are people out there who would pay good for that knowledge as long as you don't have it, so why should they tell you? ;-) Best regards Eilert Austin Lesea schrieb: > backhus, > > That is something that we thought about. But, really what we talking > about is providing access to the crypto-engine through the general > interconnect, and control of that engine. > > It was considered that anything we do in this regard would have to be > completely and thoroughly tested so as not to be a back door, and > compromise security. > > It wasn't worth the work to have to prove we did not break something.Article: 104257
Phil Hays wrote: > As an example, suppose I was designing for 4-input LUTs, and I had two > input states, and two condition bits for a state. One-hot is assumed. > The input logic to state3 would be: > > state3 :: (state1, state2, (state3, c1, c2)) > > This is five inputs, which would require two levels of logic with > 4-LUTs. I could improve speed by splitting state 3 into two states, > state3a and state3b: > > state3a :: (state1, (state3a, c1, c2)) > state3b :: (state2, (state3b, c1, c2)) > > Any output equations and next state equations would have state3 > replaced with (state3a or state3b). A "state minimization feature" > would collapse these states back into state3, making the FPGA run > slower. Since One-Hot is assumed, there probably isn't a significant time penalty for carefully packing a slice using an F5 mux to build a 5-input LUT, or using the carry mux and Cin for some cases, both of which avoid the routing/switch box delays, with a small mux delay instead. If only the tools did that automatically for performance, or allowed it easily by hand.Article: 104258
Hi, Austin Lesea schrieb: > Just don't go advertising them to be more than they really are: a > convenient way to make it cost at least $5,000 to find the key. Is it that cheap today to open the die and observe the fuses? I have no idea, if (and how) Altera protected the key fuses against optical inspection of die cuts. But If your right, it would be very cheap to reengineer most Asics. BTW Am I right, that if I use a Xilinx with security inside a equipment, the chip could be highjacked (Chipmodded) by just removing the power supply of the keys and applieing a new bitstream? Which means the bitstream itself may be protected, but not the chip? Why did nobody combine software and fuse based technologies? It would be sufficient to have 128 bit (with secure algorithms) in SW and 128 in fuses. bye ThomasArticle: 104259
Pravin G schrieb: > Hi all, > > We want to implement a design in a spartan - 3. and select a device (in terms of system gates) at the initial stage of the project. > Is there any method for this. Ask the designer who will implement the design for an estimate. (If s/he has a boss, ask the boss) There really is no other way because it is not uncommon to have a factor of 10 in LUT count between FPGA designers. For example I just saw a design that used 90% of the area for a huge mux that allowed to read back all parameter registers. Using a shift register instead or getting rid of the readback requirement would improve area 10x. Sometimes there are obvious limits as Peter suggested: The number of I/Os, number of BRAMs, number of multipliers. But LUT count is impossible to predict without knowing the engineer. Kolja SulimmaArticle: 104260
Vivek We have some V2Pro designs that we have done that might be adapted and some plans on the V4 side that might be close to what you want. Not for public consumption at present but we can talk about what you need offline. Use any of the public email addresses on our website and make it for my attention and it will be passed to me. John Adair Enterpoint Ltd. - Home of Hollybush1. The PC104+ Solution. http://www.enterpoint.co.uk "Vivek Menon" <vivek.menon79@gmail.com> wrote in message news:1150913001.916921.156290@m73g2000cwd.googlegroups.com... > Hi, > I am searching for a detachable Virtex-II Pro or a Virtex-4 FX series > FPGA with rocket I?O capability. I do not want a huge evaluation board. > After I program the FPGA and the Rocket I/O, I need to detach the > board/mini-module. > Please let me know if you are aware of such mini-modules. > Thanks, > Vivek >Article: 104261
Hi All, My Goal is to use some librarys I have for Xilinx ISE 8.1 with Cadence. The way I thought of going about this is to *somehow* export the library to EDIF and then import that EDIF to Cadence, having them in EDIF will also hopefully solve this headache if I need to use them somewhere else at a different time. Has anyone done something similar to this before, or could anyone give me some pointers. I have little to no experiance at this, I was looking at possibly writing a perl script to read from the library and create the EDIFs. Thanks DougArticle: 104262
Hi, I am using the chipshop first time. I am using this one to debug my ddr controller in the v4 board. When i tried to route the data bus of ddr through the chipscope it generated ILA and inserted into my design but when i try to implement(map and par) xilinx ise showing error that the bidirectional port is being driven by some buffer by the chipscope module.the ddr bus are bidirectional so what changes needed in the chipscope setting to route my bidirectional port. regard subinArticle: 104263
Hi All, I just needed to program and debug my Altera FPGAs remotely. So I've prepared a small script on the remote machine: #!/bin/bash killall jtagd killall jtagd sleep 1 echo I have killed the old server export PATH=/home/me/quartus/bin:$PATH sudo rmmod ppdev echo I have removed the ppdev driver sleep 2 sudo modprobe ppdev echo I've lodaed the ppdev driver sleep 2 jtagd echo I have run jtagd server sleep 4 echo Started server configuration jtagconfig --remove 1 jtagconfig --add byteblastermv lpt1 jtagconfig --enableremote my_password jtagconfig Unfortunately the last command outputs the following: 1) ByteBlasterMV [/dev/parport0] 0100A0DD ! 020030DD ! 0100A0DD ! 020030DD ! Captured DR after reset = (0100A0DD020030DD0100A0DD020030DD) [128] Captured IR after reset = (5455554555) [40] So the jtagd does not recognize the chips (probably it doesn't read the pgm_parts.txt file, when run in this mode) However the server is running and ready: $netstat -l -t -p Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name [...] tcp 0 0 *:1309 *:* LISTEN 5444/jtagd [...] When I run the jtagd with --user-start option (adding it in the proper line of the script), it correctly recognizes chips, but fails to enable remote access: Started server configuration Error (Unknown error) when setting password 1) ByteBlasterMV [/dev/parport0] 0100A0DD EPC16/4/8 020030DD EP1S25/_HARDCOPY_FPGA_PROTOTYPE 0100A0DD EPC16/4/8 020030DD EP1S25/_HARDCOPY_FPGA_PROTOTYPE now netstat returns: $ netstat -l -t -p Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name [...] tcp 0 0 localhost:1309 *:* LISTEN 5767/jtagd [...] So jtagd listens only on the local machine. I tried to use ssh tunneling: ssh user@remote.machine -L1309:localhost:1309 But then the local jtagd may not start, as 1309 port is occupied. Finally I've the solution, using the extended version of SSH tunneling: ssh user@remote.machine -Laaa.bbb.ccc.ddd:1309:localhost:1309 (aaa.bbb.ccc.ddd - is the IP of my local machine, you can also create a fake additional IP with "ifconfig eth0:1 aaa.bbb.ccc.ddd up" using an unrouteable IP) Now the local jtagd connects to the localhost 1309 port, while tunneled remote jtagd listens on the aaa.bbb.ccc.ddd 1309 port. If you now locally run $ qrt_path/jtagd --user-start $ qrt_path/jtagconfig --addserver aaa.bbb.ccc.ddd e ("e" is a dummy password, not used, but needed) $ qrt_path/bin/jtagconfig 1) ByteBlasterMV on aaa.bbb.ccc.ddd [/dev/parport0] 0100A0DD EPC16/4/8 020030DD EP1S25/_HARDCOPY_FPGA_PROTOTYPE 0100A0DD EPC16/4/8 020030DD EP1S25/_HARDCOPY_FPGA_PROTOTYPE BTW This method should provide better security. I know nothing about the security of Altera remote jtagd protocol. It seems (from ethereal dumps) that it doesn't send passwords as plaintext, but I don't know how easy it is to break. In the above method you connect via the proven SSH protocol. So my final script "~/jtagstart" looks like this: #!/bin/bash killall jtagd killall jtagd sleep 1 echo I have killed the old server export PATH=/home/me/quartus/bin:$PATH sudo rmmod ppdev echo I have removed the ppdev driver sleep 2 sudo modprobe ppdev echo I have lodaed the ppdev driver sleep 2 jtagd --user-start echo I have run jtagd server sleep 4 echo Started server configuration jtagconfig --remove 1 jtagconfig --add byteblastermv lpt1 jtagconfig You should run on the local machine: $ssh me@remote.machine -Laaa.bbb.ccc.ddd:1309:localhost:1309 /home/me/jtagstart and then locally: $ qrt_path/jtagd --user-start $ qrt_path/jtagconfig --addserver aaa.bbb.ccc.ddd e Works like a dream :-). HTH & Regards Wojtek ZabolotnyArticle: 104264
Hello Siva, Tank you very much for your help. I managed to solve the problem with the help Rob Barris, who reply to the same message posted elsewhere. It turns out that the origin of the problem was the unconnected nets. After I connected all the nets to pins on the FPGA, the problem disappeared. Regards MarianoArticle: 104265
Fred schrieb: > Does anyone know of any software, ideally freeware, which can use the above > JTAG interfaces to exercise other JTAG interfaces on non FPGA devices? > > In my case I'd like to read the state of pins on an unrelated device. There is JTAG code in the not very active open source project "open wince". KoljaArticle: 104266
In the script from the previous post, when script finishes, the tunnel disappears and connection is broken :-(. So you either must ssh to the command prompt: $ssh my@remote.machine -Laaa.bbb.ccc.ddd:1309:localhost:1309 and then run the jtagstart manually: $./jtagstart The session will be active until you logout with exit. Or you should add a never ending loop at the end of the script, e.g. like this: while true; do echo tunnel active sleep 1m done -- Sorry for confusion, wojtekArticle: 104267
Eric wrote: > > They created it this afternoon; it should be up! I'm going to import > the original code into svn right now... > ...Eric Eric, Great... I saw the version you uploaded is slightly different than the version I worked on! I'll check how my changes apply to your version and after i'll send you! bye SandroArticle: 104268
how can i do to make it work with this old devices?experimenting with an old board i need to add xcs30 and i need a tool(older impact versions) for programming it,i guess that treatin it as an xc2s30 is not the right way to proceed Thanks to everyone DiegoArticle: 104269
Diego, these parts are dinosaurs, the last version of the tools to support 4K/spartan was ISE4.1 (maybe 4.2 but I'm not sure) you can still download tools from www.xilinx.com for old parts. I don't have a link handy, but is straightforward to find them on xilinx.com Aurash blisca wrote: >how can i do to make it work with this old devices?experimenting with an old >board i need to add xcs30 and i need a tool(older impact versions) for >programming it,i guess that treatin it as an xc2s30 is not the right way to >proceed >Thanks to everyone >Diego > > > > -- __ / /\/\ Aurelian Lazarut \ \ / System Verification Engineer / / \ Xilinx Ireland \_\/\/ phone: 353 01 4032639 fax: 353 01 4640324Article: 104270
Hello, After reading up the Aurora user guide and the different replies for this topic, I have a couple of questions: Assume I have two V2 Pro or V4 boards with 8 Rocket I/O outputs each. I have 4 Rocket I/Os (TX side=4 X 2.5 or 3.125 Gbits/sec) on Board 1 connected to 4 Rocket I/Os (RX side=4 X 2.5 or 3.125 Gbits/sec) on Board2. I have also implemented Aurora protocol. Now is it possible to enable channel bonding and take the output from the 5th Rocket I/O (TX side) on Board 2 and safely assume that it transmits 10 GB/sec. If what I have described holds true, the same concept can be used to combine the data on 4 X 10 Gb/s(which cannot be achieved now). Also, if anyone has suggestions on what specific parameters are to be modified in Aurora, please let me know. I think the three parameters that need to be changed are: INST aurora_module_i_1/lane_0_mgt_i CHAN_BOND_MODE = OFF; INST aurora_module_i_1/lane_0_mgt_i CHAN_BOND_ONE_SHOT = FALSE; INST aurora_module_i_1/lane_0_mgt_i CHAN_BOND_SEQ_1_1 = 00101111100; INST aurora_module_i_1/lane_0_mgt_i SERDES_10B = FALSE; I think the CHAN_BOND_MODE should be TRUE for the 5th Rocket I/O on Board 2(cos that would be the master with the other 4 configured as slaves). What should be the values for ENHANSYNC and the remaining attributes?? Thanks in advance, Vivek Peter Alfke wrote: > Yes, we support "channel bonding", which means we can use four > transceivers, each at 2.5/3.125 Gbps to operate with a throughput of 10 > Gbps. But it involve four transceivers and four physical connections. > I think Vivek's question was whether the four channels can be combined > to one. > Physical the answer is "NO". Virtual the answer is "YES", through > channel bonding. > Peter Alfke > > ========== > Falk Brunner wrote: > > Vivek Menon schrieb: > > > Hi Sean, > > > Can you please elaborate on Virtex-II Pro X and the technique to > > > combine 4x2.5 Gbit/sec into 10 Gbit/sec? > > > > Very simple. The key word is channel bonding. This is done every day in > > the 10G XAUI interfaces. So the idea is not new and it is doable without > > too much trouble. > > > > Regards > > Falk > > > > P.S. The 4x2.5G is 4x3.125G due to 8B10B coding.Article: 104271
Hi, since the license of the "original" uart-port is expired, I'm trying to use the uartlite ip core with u-boot and it won't work. First of all the board boots PPCBoot Lite fine so I can load the compiled u-boot.bin via loadb into the memory (adress 08000000). Finally I try to boot U-boot with go 08000100 and it won't work and the board hangs. How I compiled u-boot: make AP1000_config make all it works with the "original" uart ip core uart16550 but not with uartlite. Can anybody give me a hint how to "patch" u-boot? Greetings KlausArticle: 104272
ISE and/or EDK ?Article: 104273
Vivek Menon schrieb: > Hello, > After reading up the Aurora user guide and the different replies for > this topic, I have a couple of questions: > Assume I have two V2 Pro or V4 boards with 8 Rocket I/O outputs each. I > have 4 Rocket I/Os (TX side=4 X 2.5 or 3.125 Gbits/sec) on Board 1 > connected to 4 Rocket I/Os (RX side=4 X 2.5 or 3.125 Gbits/sec) on > Board2. I have also implemented Aurora protocol. Now is it possible to > enable channel bonding and take the output from the 5th Rocket I/O (TX > side) on Board 2 and safely assume that it transmits 10 GB/sec. If the MGTs would support 10G. > If what I have described holds true, the same concept can be used to > combine the data on 4 X 10 Gb/s(which cannot be achieved now). Since you dont have a 40G MGT. But you can do channel bonding using 16 MGTs @ 3.125 Gbit/s to form al logical 40G data channel. Regards FalkArticle: 104274
Thank you very much,i guess this is the right tool,even older than i thought , http://www.xilinx.com/webpack/classics/spartan_4k/index.htm the fact is that i have to practice on surplus boards.... i ordered today an educational board from Digilent http://www.digilentinc.com/ many thanks Aurelian Lazarut <aurash@xilinx.com> wrote in message e7e4um$8oc3@cliff.xsj.xilinx.com... > Diego, > these parts are dinosaurs, the last version of the tools to support > 4K/spartan was ISE4.1 (maybe 4.2 but I'm not sure) you can still > download tools from www.xilinx.com for old parts. > I don't have a link handy, but is straightforward to find them on xilinx.com > Aurash > blisca wrote: > > >how can i do to make it work with this old devices?experimenting with an old > >board i need to add xcs30 and i need a tool(older impact versions) for > >programming it,i guess that treatin it as an xc2s30 is not the right way to > >proceed > >Thanks to everyone > >Diego > > > > > > > > > > > -- > __ > / /\/\ Aurelian Lazarut > \ \ / System Verification Engineer > / / \ Xilinx Ireland > \_\/\/ > > phone: 353 01 4032639 > fax: 353 01 4640324 > >
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