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
Michael Barr wrote: > I've just (finally) gotten around to posting a three-part article I > wrote about a year ago on computing checksums and CRCs and the source > code that goes with it. The URLs of the three articles are: > > http://www.netrino.com/Connecting/1999-11/ > http://www.netrino.com/Connecting/1999-12/ > http://www.netrino.com/Connecting/2000-01/ > > If you're new to the subject, I recommend reading them in order. If, > however, you are familiar with the theory of checksums and CRCs and just > want to get directly into the CRC implementation, you can just read the > third. > > A ZIP file containing just the source code for the CRC routines is at: > > http://www.netrino.com/Connecting/2000-01/crc.zip > > The code is placed into the public domain and may be used for any purpose > public or private. There are some limitations in its use (it's written > in C), but the underlying algorithms are solid and could be ported to the > assembly language (or hardware) of your choice if necessary. > > Enjoy, > Michael Barr Hi Michael. Question for you: Have you ever come across a method in a CRC of removing a byte, word, or long from the block of data, changing it (on purpose) and putting it back into the block of data (in its same location) and then modifying the CRC and if the CRC is checked it will check out OK? That is, I don't want to have to modify the data location and then recalculate the CRC on the whole block of data, but would like to have the CRC verify. The data is code being run in 'real time' and because it was written for one processor but being run on a new processor (the code is down loaded into the new processor's address space) and before the data/code is run, it is patched to address some instruction differences between teh original CPU target vs. the new one. Right now I'm patching the code/data and have disabled CRC checking (because I know it will fail and cause the system to halt any further processing) but I would like to (if possible) re-enable CRC checking (to catch unententional changes to the code/data) but have the checking insensitive to intentional changes to a few locations in the code/data. I know I know. No matter how I put it, it sounds goofy... Sincerely, MarcW. (And it is.)Article: 27676
Donald Gillies wrote: > Michael Barr <mbarr@netrino.com> writes: > > >This is a cryptographically signed message in MIME format. > > >--------------ms4B962FB2094CDF47ED9278AE > >Content-Type: text/plain; charset=us-ascii > >Content-Transfer-Encoding: 7bit > > >I've just (finally) gotten around to posting a three-part article I > >wrote about a year ago on computing checksums and CRCs and the source > >code that goes with it. The URLs of the three articles are: > > > http://www.netrino.com/Connecting/1999-11/ > > http://www.netrino.com/Connecting/1999-12/ > > http://www.netrino.com/Connecting/2000-01/ > > >If you're new to the subject, I recommend reading them in order. If, > >however, you are familiar with the theory of checksums and CRCs and just > >want to get directly into the CRC implementation, you can just read the > >third. > > First of all, the aforementioned piece of code appears to be > needlessly slow; the reflect() function will just nuke performance. > To avoid nuking performance, nay, if you just want a piece of code > that will BOMB performance, insteead of NUKING performance, see the > fast table-based lookup algorithm in the appendix for the PPP protocol > spec: > > http://www.cis.ohio-state.edu/htbin/rfc/rfc1331.html > > However, even this table lookup-based CRC algorithm will BOMB > performance, and if you want a decent checksum that won't destroy your > CPU, and have any freedom of design choice, > > YOU SHOULD DEFINITELY PICK A NON-GENERATOR POLYNOMIAL CHECKSUM. > > It just turns out that the typical 8-bit generator polynomial > checksum, even with a table lookup implementation, requires 5 memory > fetches to checksum each 32-bit word. Morevoer, the algebraic > instructions to do this checksum HAVE NO DATAFLOW PARALLELISM, meaning > a superscalar processor cannot achieve ANY speedup, and so you > gigahertz pentium will run the checksum loop like an 8080 at 50 Mhz. > > If you compare this to the Internet's TCP/IP CRC (1's complement > addition - a single memory fetch & add), you'll understand why we're > throwing out telecom networks and installing Internet networks all > over the earth... :) :) > > While it a 1's complement CRC-32 will not detect byte reordering, > there are similar CRC's (Fletcher's Algorithm) that still run 10x > faster than any generator polynomial checksum, and they detect byte > reordering. > > Don Gillies - gillies@netapp.com - Network Appliance, Inc. > Adjunct Professor of Computer Engineering, UBC, Vancouver BC Canada V6T 1Z4 > http://www.ee.ubc.ca/home/staff/faculty/gillies/etc/www/index.html That Fletcher checksum is limited to (fairly) small blocks of data, I think around 8191 bits (that's bits, not bytes). Sincerely, MarcW.Article: 27677
Phil James-Roxby <phil.james-roxby@xilinx.com> writes: > You could use JBits to do this. On the main Xilinx web site, go to > products -> System solutions -> Xilinx Online and there is a little > description of JBits. Sounds nifty. It says "JBits will be available over the web in 1Q99." But I can't seem to actually find it. Any hints?Article: 27678
Well, you can run a CRC forwards and reverse. The reverse is actually a different polynomial. See DSP Guru for the trick, but basically you just take the (order - power) for each term to get the other polynomial. You should be able to use this info to gen a CRC patch - I dunno how though, but I think it would be workable. If you get completely stuck, drop me some mail, I know a few tricks. Cheers, Herman http://www.AerospaceSoftware.com Marc Warden wrote: > > Michael Barr wrote: > > > I've just (finally) gotten around to posting a three-part article I > > wrote about a year ago on computing checksums and CRCs and the source > > code that goes with it. The URLs of the three articles are: > > > > http://www.netrino.com/Connecting/1999-11/ > > http://www.netrino.com/Connecting/1999-12/ > > http://www.netrino.com/Connecting/2000-01/ > > > > If you're new to the subject, I recommend reading them in order. If, > > however, you are familiar with the theory of checksums and CRCs and just > > want to get directly into the CRC implementation, you can just read the > > third. > > > > A ZIP file containing just the source code for the CRC routines is at: > > > > http://www.netrino.com/Connecting/2000-01/crc.zip > > > > The code is placed into the public domain and may be used for any purpose > > public or private. There are some limitations in its use (it's written > > in C), but the underlying algorithms are solid and could be ported to the > > assembly language (or hardware) of your choice if necessary. > > > > Enjoy, > > Michael Barr > > Hi Michael. > > Question for you: Have you ever come across a method in a CRC of removing a > byte, word, or long from the block of data, changing it (on purpose) and > putting it back into the block of data (in its same location) and then > modifying the CRC and if the CRC is checked it will check out OK? > > That is, I don't want to have to modify the data location and then > recalculate the CRC on the whole block of data, but would like to have the > CRC verify. > > The data is code being run in 'real time' and because it was written for one > processor but being run on a new processor (the code is down loaded into the > new processor's address space) and before the data/code is run, it is patched > to address some instruction differences between teh original CPU target vs. > the new one. > > Right now I'm patching the code/data and have disabled CRC checking (because > I know it will fail and cause the system to halt any further processing) but > I would like to (if possible) re-enable CRC checking (to catch unententional > changes to the code/data) but have the checking insensitive to intentional > changes to a few locations in the code/data. > > I know I know. No matter how I put it, it sounds goofy... > > Sincerely, > > MarcW. (And it is.)Article: 27679
"InGenius Engineering" <Ellen@ingeniusengineering.com> wrote: >Hi there, > >My name is Ellen Ann Nichol and I am a recruiter...before you spam me..I am >not hear to use this posting board to recruit FPGA people. Since I have a ^^^^ >dire need for FPGA people, I only wanted to know if people could send me >suggestions on how to recruit you guys. Call it a survey if you like. If >you were looking for a job where would you look? Do you go to certain web >sites? That sort of thing. Any tips would be greatly appreciated. > >ellen@ingeniusengineering.com > > I always see "great written and verbal communication skills" in engineering job descriptions. Isn't this also requirement for recruiters? Muzaffer FPGA DSP Consulting http://www.dspia.comArticle: 27680
If you are using the DLLs in VIrtex, you can't really stop the clock because it makes the DLL lose lock. In fact there is a configuration mode that will hold GSR on until the DLL locks. WIthout the DLL, you gt pretty poor clock to out timing, which tends to make the DLL a requirement for moderate to high clock rates. pbmueller@my-deja.com wrote: > > What I have done in a Lucent (OR3T80) design is to disable the clock as long > as the GSR is asserted (and a few cycles longer). This should also be > possible in the virtex parts, but... -The BUFG primitive does not show the CE > (Clock Enable) input of the elemet, that can be seen in the FPGAEditor -If > one creates a macro that allows to access the CE signal, this only works when > you do not use a DLL (This is because of a bug of the Alliance 3.2i software, > that does not handel the XIL_MAP_ALLOW_ANY_DLL_INPUT variable as it > should...and stops with an error) > > Patrick > > In article <3A27369A.BDF76883@andraka.com>, > Ray Andraka <ray@andraka.com> wrote: > > Use GSR to initialize all your FF's, but then have only one point in your design > > depend on when the GSR is released. For example, use a synchronized version of > > GSR to hold a master state machine in reset after GSR is released, then after > > GSR is safely gone away let the thing start going. > > Sent via Deja.com http://www.deja.com/ > Before you buy. -- -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: 27681
Hi, I am looking for a synthesizable non-commercial / open source pcmcia host controller for quick prototyping purpose. Any suggestions. Regards, ChrisArticle: 27682
--------------64DBD04C3BE1155F277E6C3D Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello Dan, Its possible to connect two DLLs in series to generate a 4x clock. Each DLL has +/- 60ps jitter and this may be accumulated or compensated across both DLLs. Just make sure the external input clock doesn't have any jitter, as this would be propagated through the DLLs. Other than this, the 4x clock will be stable and does not loose lock unless one of the DLL is reset or the input clock frequency changes. You can refer XAPP132: Using the Virtex Delay-Locked Loop available at http://support.xilinx.com/apps/virtexapp.htm This app. note explains details on generating 4x clock with two DLLs and other DLL applications. Also refer Virtex datasheet for DLL specs. -Vikram Xilinx Applications Dan wrote: > Hello, > > Is is possible to drive the input of a DLL with the output of another DLL ? > > I need to multiply by 4. A DLL will only multiply by 2. > > Will the resultant 4X clock have poor characteristics if I do this? > (unstability/ bad jitter/ loss of lock ......) ? > > Also, I may want to use tw DLL to get multiply by 1.3333 ( 1st DLL 2X, 2nd > DLL /1.5) > > Sincerely > Daniel DeConinck > High Res Technologies, Inc. --------------64DBD04C3BE1155F277E6C3D Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit <!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html> Hello Dan, <p>Its possible to connect two DLLs in series to generate a 4x clock. Each DLL has +/- 60ps jitter and this may be accumulated or compensated across both DLLs. Just make sure the external input clock doesn't have any jitter, as this would be propagated through the DLLs. Other than this, the 4x clock will be stable and does not loose lock unless one of the DLL is reset or the input clock frequency changes. <p>You can refer <a href="http://support.xilinx.com/xapp/xapp132.pdf">XAPP132: Using the Virtex Delay-Locked Loop</a> available at <br><A HREF="http://support.xilinx.com/apps/virtexapp.htm">http://support.xilinx.com/apps/virtexapp.htm</A> <p>This app. note explains details on generating 4x clock with two DLLs and other DLL applications. Also refer Virtex datasheet for DLL specs. <p>-Vikram <br>Xilinx Applications <br> <br> <br> <br> <p>Dan wrote: <blockquote TYPE=CITE>Hello, <p>Is is possible to drive the input of a DLL with the output of another DLL ? <p>I need to multiply by 4. A DLL will only multiply by 2. <p>Will the resultant 4X clock have poor characteristics if I do this? <br>(unstability/ bad jitter/ loss of lock ......) ? <p>Also, I may want to use tw DLL to get multiply by 1.3333 ( 1st DLL 2X, 2nd <br>DLL /1.5) <p>Sincerely <br>Daniel DeConinck <br>High Res Technologies, Inc.</blockquote> </html> --------------64DBD04C3BE1155F277E6C3D--Article: 27683
Hello: But,the type of software i use is Foundation2.1,after i synthesis the project on PC ,it will creat a "time_sim.v" file.I translate it to Sun Sloaris Workstation .Finally,i use NC-VERILOG to simulate it.The questio i mentioned will occrue.How do i? Thanks a lotArticle: 27684
FPGAs are the subject of my latest column on adaptive systems. This column appears every two weeks in eChips’ ChipCenter. The current column appears at the second URL below. Click on the “more” link to see the entire column. The first URL below is my general site on adaptive automation. Best, Peter. Peter G. Raeth Computer Engineer, Specialist in Adaptive Systems peter_raeth@juno.com http://www.geocities.com/siliconvalley/lakes/6007 http://www.chipcenter.com/eexpert/praeth/main.html Sent via Deja.com http://www.deja.com/ Before you buy.Article: 27685
mmeraliuk@yahoo.co.uk wrote: > Does anyone know of a way to prevent Xilinx A2.1i PAR from routing > through a particular CLB's routing resources? > > Cheers, > matthew. > > Sent via Deja.com http://www.deja.com/ > Before you buy. I don't know if the tool still uses the routing resource of a definite area, but it is possible to hinder the placement process of PAR from using a definite area to place the logic: Use PROHIBIT. Look at the manuals how to use it efficiently. By the way, switch to A3.x ASAP, ow you won't get much help from around!! Utku -- I feel better than James Brown.Article: 27686
Eric Smith <eric-no-spam-for-me@brouhaha.com> writes: > Phil James-Roxby <phil.james-roxby@xilinx.com> writes: > > You could use JBits to do this. On the main Xilinx web site, go to > > products -> System solutions -> Xilinx Online and there is a little > > description of JBits. The page I visited (did search for jbits from search page) was: http://www.xilinx.com/products/software/jbits/index.htm JBits SDK > Sounds nifty. Is nifty. As far as what I have read in its documentation, since installing it last wednesday. Now I just need time to use it ... >It says "JBits will be available over the web in 1Q99." > But I can't seem to actually find it. Any hints? Mail to the address given on the page (JBits@xilinx.com). They then send you an URL (actually 3 URLs, Windows NT, Solaris, Linux) plus password. Then download (9MByte), install and have fun. Greetings from another a.f.c-er. Looks like JBits and an XCV300 are going to be the base of some historic computer cloning... -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Nerd, Geek, Hacker, Unix Guru, Sysadmin, Roleplayer, LARPer, MysticArticle: 27687
Hi, I haven't used Foundation Series, but I have used the latter one. chsw wrote: > > Hello: > But,the type of software i use is > Foundation2.1,after i synthesis the project on PC ,it will creat a "time_sim.v" file.I translate it to Sun Sloaris Workstation . What do you actually mean by "translate" - guess you do an FTP or so (to "transfer" your file from PC to UNIX) - Am I correct? If so "transfer" also the file named " $XILINX/verilog/src/glbl.v " to UNIX (and remove Control-M s if reqd) and try what Paulo suggested i.e. ncvlog -update $XILINX/verilog/src/glbl.v <testfixture>.v time_sim.v ncelab -messages -autosdf testfixture_name glbl ncsim -messages testfixture_name HTH, Srini -- Srinivasan Venkataramanan (Srini) ASIC Design Engineer, Chennai (Madras), IndiaArticle: 27688
Hello: what is the "testfixture"? I try to do it according to what you said,but the error occrued: " ncelab:*W,SDFNCS:Not a compiled SDF file :time_sim.sdf-compiling it....... ncelab:*E,CUVMOT(./pos_bench.v,15|25):illegal output ports specification. test_convert my (reop,tenb,rx_err........tx_pa0, why? how do i? is it right? "ncvlog -update /opt/cadence/src/glbl.v /opt/cadence/src/simprims/*.v pos_bench.v time_sim.v -line -mess ncelab pos_bench glbl -autosdf -access+rcw-mess ncsim pos_bench -gui -mess & Thanks a lot!Article: 27689
I am starting a new board design where I wanted to use the SpartanII FPGAs. But I am not sure if I should continue with these parts or change to another. According to a January press release, the SpartanII chips were supposed to be in full production "within the current quarter". Here it is nearly a year later and they are still not listed as being readily available at distributors. The latest data sheet (Nov 2, 2000) is still labeled as "Preliminary". The other problem I have is with the high startup current these parts require. I need to be able to sell these boards for an industrial (-40C to 85C) temp range. At startup in this range, the four parts on this board will require 2 amps from the 2.5 volt supply. Otherwise this supply only needs to be built to handle 1 amp including power to spare. I have seen several posts here that seem to indicate that the startup current requirement may be spec'd a little more tightly when these parts are in fully qualified. Does anyone have an idea of when these parts will be fully qualified and in full production? I have been burned before trying to use parts that are new and not available. But these parts are "just what the doctor ordered" other than these two issues. I know that Lucent has no plans for product to compete with SpartanII at the low end. Does anyone know if Altera is introducing a similar low cost, high density family of parts? -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. 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 URL http://www.arius.comArticle: 27690
InGenius Engineering wrote: > Hi there, > > My name is Ellen Ann Nichol and I am a recruiter...before you spam me..I am > not hear to use this posting board to recruit FPGA people. Since I have a > dire need for FPGA people, I only wanted to know if people could send me > suggestions on how to recruit you guys. Call it a survey if you like. If > you were looking for a job where would you look? Do you go to certain web > sites? That sort of thing. Any tips would be greatly appreciated. > > ellen@ingeniusengineering.com Tip #1 :- Avoid all the usual BS words like ``dynamic'', ``innovative'', ``leading-edge'', etc. etc. These are usually a substitute for not being able to name the company. If you can't name it just say so. Just tell us what the job's about I we'll work out whether its ``dynamic'', ``exciting'', etc enough for us. Tip #2 :- Don't be coy about the numbers of $$$, £££, YYY, EEE etc. Tip #3 :- Remember that if you do try & recruit via the NGs then you're likely to be talking to some seriously battle-hardened digital designers who really don't appreciate being talked down to. If you want recent examples of what happens if you don't follow this rule look up the ``Very Lucrative FPGA jobs'' and ``Long Island Verilog'' threads on dejanews.Article: 27691
Rick Collins wrote: > > > I know that Lucent has no plans for product to compete with SpartanII at > the low end. Does anyone know if Altera is introducing a similar low > cost, high density family of parts? > > -- > > Rick "rickman" Collins Rick, I haven't seen a data sheet for one, but I believe the Altera Acex family are targeted at this market. Might be worth a look. Nial.Article: 27692
On Sun, 03 Dec 2000 04:31:22 -0500, Rick Collins <spamgoeshere4@yahoo.com> wrote: >I am starting a new board design where I wanted to use the SpartanII >FPGAs. But I am not sure if I should continue with these parts or change >to another. According to a January press release, the SpartanII chips >were supposed to be in full production "within the current quarter". >Here it is nearly a year later and they are still not listed as being >readily available at distributors. The latest data sheet (Nov 2, 2000) >is still labeled as "Preliminary". > >The other problem I have is with the high startup current these parts >require. I need to be able to sell these boards for an industrial (-40C >to 85C) temp range. At startup in this range, the four parts on this >board will require 2 amps from the 2.5 volt supply. Otherwise this >supply only needs to be built to handle 1 amp including power to spare. >I have seen several posts here that seem to indicate that the startup >current requirement may be spec'd a little more tightly when these parts >are in fully qualified. > >Does anyone have an idea of when these parts will be fully qualified and >in full production? I have been burned before trying to use parts that >are new and not available. But these parts are "just what the doctor >ordered" other than these two issues. > >I know that Lucent has no plans for product to compete with SpartanII at >the low end. Does anyone know if Altera is introducing a similar low >cost, high density family of parts? I did a PCI design with Spartan II and had similar issues. Actually Spartan II is available but lead time is a little bit long. My client managed to get 5 samples in two weeks and ordered 200 to be delivered 2 months. This was 3 months ago. Now they are going to production and expecting delivery of more parts. I think just a little patience after placing your order is needed here. Muzaffer FPGA DSP Consulting http://www.dspia.comArticle: 27693
>It is true that Xilinx has had to shorten the lifecycle of the 'legacy' Philips >parts due to lack of fab capacity. As you probably know, Xilinx has strategic >partnerships with a couple of fabs, but unfortunately not with the fabs where >these devices are manufactured. Which Xilinx knew when they took over the line... >Hence the end of life notice. Which should have been issued the day Xilinx took it over, frankly. >It has always >been in the plans to move the newest CoolRunner XPLA3 family to a strategic fab >partner, and that is in the works now. Except there will be no continuity of the PxZ22V10 which is a very useful and well priced device. Bye to Xilinx - I am designing them out as I write. Fortunately I have other options. Peter. -- Return address is invalid to help stop junk mail. E-mail replies to zX80@digiYserve.com but remove the X and the Y. Please do NOT copy usenet posts to email - it is NOT necessary.Article: 27694
Hi, I am working on a new design , using XCV400-BG560. How can I find out (except looking at the Pin-out symbol diagram) which I/O pin belongs to each of the banks ? Is there a source for Orcad symbols for Virtex family parts or do I need to generate my own ?. Thanks, Rotem. Sent via Deja.com http://www.deja.com/ Before you buy.Article: 27695
"Rick Filipkiewicz" <rick@algor.co.uk> wrote in message news:3A2A1B70.4E87C014@algor.co.uk... > Tip #2 :- > > Don't be coy about the numbers of $$$, £££, YYY, EEE etc. I wouldn't discuss $$$ on the Internet. Everyone is in a different situation and the money will be different based on several factors, including the actual job and expertise involved. If you discuss money and the money is good, expect the lowly types to respond immediately wanting that rate. If the money is bad, expect no one to respond. In general, I feel that the parties should "feel" each other out before discussing money. If you throw money on the Internet, expect used car salesmen that surf the Internet to suddenly change careers and respond. -Simon Ramirez, Consultant Synchronous Design, Inc.Article: 27696
Nial Stewart wrote: > > Rick Collins wrote: > > > > > > I know that Lucent has no plans for product to compete with SpartanII at > > the low end. Does anyone know if Altera is introducing a similar low > > cost, high density family of parts? > > > > -- > > > > Rick "rickman" Collins > > Rick, > > I haven't seen a data sheet for one, but I believe the Altera Acex > family are targeted at this market. > > Might be worth a look. > > Nial. Thanks Nial, Yes, I found that today. I have never used Altera parts before, but I took a look today and found that the Acex family is just what I wanted. Altera uses a different approach to compatibility. Rather than put the different chips in the same package, they also use footprint compatible FBGAs which allow you to put down the largest footprint you think you may need. You can then use any of the smaller FBGAs in the same footprint. I have not checked the pinouts in detail, but they seem to advertise this pretty loudly. The pricing looks good. The Acex chips similar to the SpartanII chips I was going to use have the same total price. But I will be looking hard at the startup current issues as well as availabilty. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. 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 URL http://www.arius.comArticle: 27697
Muzaffer Kal wrote: > > On Sun, 03 Dec 2000 04:31:22 -0500, Rick Collins > <spamgoeshere4@yahoo.com> wrote: > >Does anyone have an idea of when these parts will be fully qualified and > >in full production? I have been burned before trying to use parts that > >are new and not available. But these parts are "just what the doctor > >ordered" other than these two issues. > > > >I know that Lucent has no plans for product to compete with SpartanII at > >the low end. Does anyone know if Altera is introducing a similar low > >cost, high density family of parts? > > I did a PCI design with Spartan II and had similar issues. Actually > Spartan II is available but lead time is a little bit long. My client > managed to get 5 samples in two weeks and ordered 200 to be delivered > 2 months. This was 3 months ago. Now they are going to production and > expecting delivery of more parts. I think just a little patience after > placing your order is needed here. > > Muzaffer > > FPGA DSP Consulting > http://www.dspia.com I can't take a chance with 2 month lead times. When I get an order for boards, I can't quote a three or four month lead time because I have to wait 2 months or more for FPGAs. The order will vanish! If the chips are being built as production parts, why are November data sheets still labeled "Preliminary"? -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. 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 URL http://www.arius.comArticle: 27698
Rick Collins wrote: > > Nial Stewart wrote: > > > > Rick Collins wrote: > > > > > > > > > I know that Lucent has no plans for product to compete with SpartanII at > > > the low end. Does anyone know if Altera is introducing a similar low > > > cost, high density family of parts? > > > > > > -- > > > > > > Rick "rickman" Collins > > > > Rick, > > > > I haven't seen a data sheet for one, but I believe the Altera Acex > > family are targeted at this market. > > > > Might be worth a look. > > > > Nial. > > Thanks Nial, > > Yes, I found that today. I have never used Altera parts before, but I > took a look today and found that the Acex family is just what I wanted. > The pricing looks good. The Acex chips similar to the SpartanII chips I > was going to use have the same total price. But I will be looking hard > at the startup current issues as well as availabilty. One thing to look out for with Altera devices/tools is unused pins. With the Xilinx tools if a pin in unused it is tri-stated, but with the Altera tools it's arbitrarily connected to an internal net and driven out. It is therefore _not_ a good idea to build a design in incremental stages incorporating more pins/functionality as you go. You must define all used pins (in your code and .acf file) right from the start then the Altera tools will tri-state them if unused. I was bitten by this when mainting someone elses design. For some reason on of the uP address input pins had been commented out of the code as it wasn't being used for address decoding. As this wasn't used the tools cnnected it to a net and drove it out, where it was driving against the uP. This worked for a while, then fell over and took ages to track down. Uncommenting the unused input fixed the problem. Nial.Article: 27699
I have a similar problem. I would like to change a table in EEPROM and still have the CRC verify correctly. It should be possible to provide an additional dummy table with some "CRC correction bytes", which would bring the CRC back to the original value, so that the table data could be modified as desired, provided that the appropriate "correction bytes" are also modified. Question is, how does one go about determining rules to generate the correction bytes, for any given table data? I assume that at least the following information would be required: - Initialisation CRC (CRC calculated to beginning of table) - Static bytes between table and dummy table, assuming dummy table does not directly follow the first table in memory. - Initial bytes in table, and dummy table Any ideas, Herman? In article <3A28514B.7F695805@AerospaceSoftware.com>, Herman Oosthuysen <aerosoft@AerospaceSoftware.com> wrote: > Well, you can run a CRC forwards and reverse. The reverse is actually a > different polynomial. See DSP Guru for the trick, but basically you > just take the (order - power) for each term to get the other > polynomial. You should be able to use this info to gen a CRC patch - I > dunno how though, but I think it would be workable. If you get > completely stuck, drop me some mail, I know a few tricks. > > Cheers, > > Herman > http://www.AerospaceSoftware.com > > Marc Warden wrote: > > > > Michael Barr wrote: > > > > > I've just (finally) gotten around to posting a three-part article I > > > wrote about a year ago on computing checksums and CRCs and the source > > > code that goes with it. The URLs of the three articles are: > > > > > > http://www.netrino.com/Connecting/1999-11/ > > > http://www.netrino.com/Connecting/1999-12/ > > > http://www.netrino.com/Connecting/2000-01/ > > > > > > If you're new to the subject, I recommend reading them in order. If, > > > however, you are familiar with the theory of checksums and CRCs and just > > > want to get directly into the CRC implementation, you can just read the > > > third. > > > > > > A ZIP file containing just the source code for the CRC routines is at: > > > > > > http://www.netrino.com/Connecting/2000-01/crc.zip > > > > > > The code is placed into the public domain and may be used for any purpose > > > public or private. There are some limitations in its use (it's written > > > in C), but the underlying algorithms are solid and could be ported to the > > > assembly language (or hardware) of your choice if necessary. > > > > > > Enjoy, > > > Michael Barr > > > > Hi Michael. > > > > Question for you: Have you ever come across a method in a CRC of removing a > > byte, word, or long from the block of data, changing it (on purpose) and > > putting it back into the block of data (in its same location) and then > > modifying the CRC and if the CRC is checked it will check out OK? > > > > That is, I don't want to have to modify the data location and then > > recalculate the CRC on the whole block of data, but would like to have the > > CRC verify. > > > > The data is code being run in 'real time' and because it was written for one > > processor but being run on a new processor (the code is down loaded into the > > new processor's address space) and before the data/code is run, it is patched > > to address some instruction differences between teh original CPU target vs. > > the new one. > > > > Right now I'm patching the code/data and have disabled CRC checking (because > > I know it will fail and cause the system to halt any further processing) but > > I would like to (if possible) re-enable CRC checking (to catch unententional > > changes to the code/data) but have the checking insensitive to intentional > > changes to a few locations in the code/data. > > > > I know I know. No matter how I put it, it sounds goofy... > > > > Sincerely, > > > > MarcW. (And it is.) > Sent via Deja.com http://www.deja.com/ Before you buy.
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