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
Hi, I am new to using lwip. I am wondering if you can advice me on the same. I need to transfer data from PC to Microblaze microprocessor on a Xilinx FPGA board and vice versa as fast as possible. i dun need any error correction or the overheads like full TCP/IP. I am currently readin this article to understand about socket programming as an initial start: http://cscene.unitycode.org/topics/unix/cs2-13.xml.html Here are the questions that I have: 1) I am stuck at the part in the link above which says hp=gethostbyname(servername);a the code above is more for developing applications on PC. I am looking for lwip drivers to run on the Microblaze (32bit CPU) on the FPGA. cos i cannot find a similar function in lwip. any idea how to resolve this? i have tried looking for equiv functions like getnameinfo(), getaddrinfo() but these functions will work. :( 2) any other useful links to read for regards to this topic? i have tried looking at http://www.xilinx.com/ise/embedded/edk91i_docs/lwip_v2_00_a.pdf what do i need to purchase to let the sockets running on both PC and FPGA communicate with each other? possible to advice me on some documents to point me in the right direction? thanks ChrisArticle: 133451
On Jun 29, 4:28=A0am, Philipp Hachtmann <hac...@hachti.de> wrote: > Hi, > > I am using an EDK-generated PPC subsystem on a ML403. > I wonder if it would be complicated to create a peripheral that has high > =A0 speed master access to the SDRAM. I would like to build a shared > memory frame buffer. Or are there any other good ways to achieve my > goal? I don't want to use the onboard SRAM. And block ram is too rare > and not enough.. > > Regards, > Philipp :-) Hi Phillip, You already have a frame buffer in EDK 10.1 as a plugin to MPMC3 memory controller. If it does not fit your requirements you can still modify it. Otherwise you can download the source from opencores.org Remember that MPMC offers very high bandwidth. Cheers, GuruArticle: 133452
On Jun 27, 7:52=A0pm, Chris Maryan <kmar...@gmail.com> wrote: > On Jun 27, 11:05=A0am, "Norman Bollmann" <wirdnichtgele...@gmx.net> > wrote: > > > Thank you very much for your feedback, so far! > see AMMASS project there was a demo > > Oh my.... I'm sorry. That was definitely a spelling mistake: it's suppo= sed > > to be "26290144 elements". Sorry 'bout that! > > > Regards > > > Norman Bollmann > > Even with 26M elements to search, as long as they're only 16bits, > that's still only 65k choices. A hash table based thing as someone > else pointed out would be a good option. > > Can you give some details about the search you need to perform? > > ChrisArticle: 133453
see AMMASS project there was a demo at IP07 grenoble http://www.design-reuse.com/ip07Article: 133454
Hi everybody, I am using PPC to configure number of LUTs using HWICAP, my question is , Is it possible to generate an interrupt internally according to a change in register value , as i want to do new configuration depending on this value? thanks FatmaArticle: 133455
On Sun, 29 Jun 2008 20:30:30 +0200, Thorsten Kiefer <webmaster@nillakaes.de> wrote: >Hi, >the following code : > >library IEEE; >use IEEE.STD_LOGIC_1164.ALL; >use IEEE.numeric_std.ALL; >... >rnext.addr <= std_logic_vector(unsigned(rreg.addr) + 1); >... > >yields the following error : >Line 140. Expression in type conversion to std_logic_vector has 3 possible >definitions in this scope, for example, UNSIGNED and UNSIGNED. > >What can I do ? Declare rnext.addr and rreg.addr as unsigned, and write rnext.addr <= rreg.addr + 1; In my experience, if you're fighting the type system, you're probably working at the wrong level of abstraction. Addresses usually have the semantics of unsigned numbers (with the possible exception of bit-twiddling address reordering for FFTs) so why not use unsigned types directly? Only when feeding the address to an external memory do I convert its type to std_logic_vector. (Where I instantiate internal memories I have to do the same; I use a wrapper for these to hide the instantiated component anyway to improve portability; the conversion happens there) - BrianArticle: 133456
On Jun 29, 11:36 pm, mahesh <mahesh...@gmail.com> wrote: > On Jun 29, 5:54 pm, timinganalyzer <timinganaly...@gmail.com> wrote: > > > > > Hello All, > > > A new version beta 0.85 is now available. The following changes and > > additions have occurred. > > > 1 Quickly add previously used Delays and Constraints from pop-up menu > > 2 The current state is inverted automatically when adding new pulse if > > the newstate in the > > toolbar is the same as the current state in the signal. > > 3 Keystroke shortcuts for all the signal state buttons in toolbar. > > Hover over button to see keystrokes. > > 4 Signals can now use text values for next state. This is useful when > > labels are needed in digital signals. > > 5 Moving text and edges by mouse drags now use undo and redo > > 6 Should work with JRE1.5 or newer > > > You can download the Free Edition now and read all about the > > TimingAnalyzer at: > > >www.timing-diagrams.com > > Hi, > I tried to install it on Vista Basic.But facing difficulty in > installing.I did install the JRE.when I run the timinganalyzer.jar exe > file It pops me a message saying main class not found exiting !! > Could u pls help me in fixing the problem ...ur help will be > appreciated > > Thanks in advance, > hesh Can you tell what version of Java you are using? I have used beta 0.85 with JRE1.5. If you bring up a dos window, enter java -jar TimingAnalyzer.jar This will report the OS, Java version, .... Thanks, DanArticle: 133457
On Jun 28, 3:08 pm, Jim Granville <no.s...@designtools.maps.co.nz> wrote: > KJ wrote: > > "General Schvantzkopf" <schvantzk...@yahoo.com> wrote in message > >news:y-2dnZJNvsTS8vvVnZ2dnUVZ_sjinZ2d@comcast.com... > > >>On Sat, 28 Jun 2008 09:00:28 +0100, Evan Lavelle wrote: > > >>>On Sat, 28 Jun 2008 09:00:03 +1200, Jim Granville > >>><no.s...@designtools.maps.co.nz> wrote: > > > <snip> > > >>It's been 30 years since I've used Karnaugh maps, they were useful when > >>the 7400 was state of the art and every gate counted. In that era logic > >>was necessarily simple so a tool that could minimize a four input equation > >>was helpful. > > > All very true, but Evan's question was explicitly targetted toward people > > who USE K-maps ("If you use K-maps, which version do you prefer..." from the > > original post) and his reasons had to do with verification not design (later > > post). > > To answer that question, the Logic minimisation I use, is that built > into the tools, which will include the tools version of K-Maps, and > others as well. ie the REPORT files are closely scrutinised. > > It is common to look at the tool output, and recode the source, and > sometimes we have taken one tools output, and pasted into another > as source code. > > Sometimes, the choices made within the tools surprises, and sometimes it > disappoints.... > > Where the tools have a blind spot is in 'node creation' to reduce > total mapped logic (at the expense of some speed). Understandable, > as that is a LOT more variables in the air. > > Sometimes the designer has to help there. > > It's a matter of matching the degrees of freedom, to the tools IQ. > > In Xilinx flows, fastest and smallest switches sometimes seem to work > backwards :) > > -jg I rarely use a K map for "logic minimization". Once I got out of school (some 22 years ago), about the only thing I used them for was for state-code assignments in state machines with unsynchronized inputs (guaranteeing gray-code state transitions on those inputs), and for P-term minimization when using PALs (choosing state assignments to minimize branches into states with bits set). With modern FPGAs in most cases, if you are designing at the k-map level, you are not taking maximum advantage of the tools. Specifying design behavior at higher levels of abstraction often leads to correspondingly higher levels of optimization. But not always... Another advantage of using higher levels of abstraction is in debugging and verification. State transition optimization to avoid an extra register for input synchronization may have been a good idea when an entire device had only 8 or 10 registers. But when extra registers are so darn cheap in an FPGA, it is foolhardy, and not nearly as easily verified by inspection as simply burning another register to synchronize the input. Andy From webmaster@nillakaes.de Mon Jun 30 08:34:14 2008 Path: flpi142.ffdc.sbc.com!flpi088.ffdc.sbc.com!prodigy.com!flpi089.ffdc.sbc.com!prodigy.net!goblin2!goblin.stu.neva.ru!feeder.erje.net!newsfeed.datemas.de!feed.cnntp.org!news.cnntp.org!not-for-mail Message-Id: <4868fcf6$0$25952$6e1ede2f@read.cnntp.org> From: Thorsten Kiefer <webmaster@nillakaes.de> Subject: Re: arithmetic problem Newsgroups: comp.arch.fpga Date: Mon, 30 Jun 2008 17:34:14 +0200 References: <4867d4d0$0$25951$6e1ede2f@read.cnntp.org> <ao3h64ln9u4fvmhlfn2v7gfuj45cntoak3@4ax.com> User-Agent: KNode/0.10.4 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7Bit Lines: 22 Organization: CNNTP NNTP-Posting-Host: ca153496.read.cnntp.org X-Trace: DXC=d9kfFF0[[I[TWP9QK:59h[WoT\PAgXa?QZ1ejXFd@BcVCWCf629N19T_ZoYESFc?9ZH?EJh[X2M]Sh6e3V^SUc8V X-Complaints-To: abuse@cnntp.org Xref: prodigy.net comp.arch.fpga:146049 X-Received-Date: Mon, 30 Jun 2008 11:34:15 EDT (flpi142.ffdc.sbc.com) > 1) Make absolutely sure that you don't have > use ieee.std_logic_arith > or > use ieee.std_logic_[un]signed > in the context clauses. I don't use them ... > > 2) Just as an experiment (I'm not recommending this for real code) > try > > rnext.addr <= std_logic_vector( > unsigned'(unsigned(rreg.addr) + 1) > ); rnext.addr <= std_logic_vector( unsigned'(unsigned(rreg.addr) + to_unsigned(18,1) ); works fine. Regards ThorstenArticle: 133458
On Jun 30, 11:34 am, Thorsten Kiefer <webmas...@nillakaes.de> wrote: > > 1) Make absolutely sure that you don't have > > use ieee.std_logic_arith > > or > > use ieee.std_logic_[un]signed > > in the context clauses. > > I don't use them ... > > > > > 2) Just as an experiment (I'm not recommending this for real code) > > try > > > rnext.addr <= std_logic_vector( > > unsigned'(unsigned(rreg.addr) + 1) > > ); > > rnext.addr <= std_logic_vector( > unsigned'(unsigned(rreg.addr) + to_unsigned(18,1) > ); > works fine. > > Regards Thorsten Do you mean this: rnext.addr <= std_logic_vector( unsigned'(unsigned(rreg.addr) + to_unsigned(1,18); to_unsigned(18,1) converts the integer 18 to unsigned which is one bit wide. I think you meant to convert the integer 1 to an unsigned which is 18 bits wide. Also, I don't think you need to cast the result of the addition to unsigned, since the result of the addition of two unsigneds is going to be an unsigned. There is no ambiguity there. Dave DaveArticle: 133459
On Jun 28, 1:42 pm, Jim Flanagan <jf...@tampaREMOVEbay.rr.com> wrote: > Guys.. > I am *brand new* to Aldec Active HDL. What few > cpld/pals that I have done have been with CUPL. > > I've started using ActiveHDL(Student Ver), writing some easy VHDL > just to get acquainted with both VHDL and the Aldec tool. > What I don't understand is the integrated implementation > portion. If I were to want to do a Altera design, for example, > do I need to have the Altera toolset installed in order to > build the jedec file or does the Aldec tool do that also? > I guess I am at a loss at the point between designing/simulating the > project and implementation with ActiveHDL. > > Help me see the light. > Thanks, > Jim You still need a separate tool to do the synthesis, whether it's Synplicity or ISE or Quartus. Then, you still need ISE or Quartus even if you used Synplicity for the synthesis, to do the final place and route, and build the file to download into the FPGA (bit or pof file). Aldec will only do simulations, but it does allow you to launch these other tools from within its GUI. You don't have to, though, you could launch them yourself if you find Aldec's design flow confusing. DaveArticle: 133460
On Jun 30, 11:06 am, Dave <dhsch...@gmail.com> wrote: > On Jun 30, 11:34 am, Thorsten Kiefer <webmas...@nillakaes.de> wrote: > > > > > > 1) Make absolutely sure that you don't have > > > use ieee.std_logic_arith > > > or > > > use ieee.std_logic_[un]signed > > > in the context clauses. > > > I don't use them ... > > > > 2) Just as an experiment (I'm not recommending this for real code) > > > try > > > > rnext.addr <= std_logic_vector( > > > unsigned'(unsigned(rreg.addr) + 1) > > > ); > > > rnext.addr <= std_logic_vector( > > unsigned'(unsigned(rreg.addr) + to_unsigned(18,1) > > ); > > works fine. > > > Regards Thorsten > > Do you mean this: > > rnext.addr <= std_logic_vector( > unsigned'(unsigned(rreg.addr) + to_unsigned(1,18); > > to_unsigned(18,1) converts the integer 18 to unsigned which is one bit > wide. I think you meant to convert the integer 1 to an unsigned which > is 18 bits wide. > > Also, I don't think you need to cast the result of the addition to > unsigned, since the result of the addition of two unsigneds is going > to be an unsigned. There is no ambiguity there. > > Dave > > Dave You also need not convert the integer to unsigned. Numeric_std "+" is already defined for unsigned and natural operands (returning unsigned). next.addr <= std_logic_vector(unsigned(rreg.addr) + 1); AndyArticle: 133461
On Jun 30, 5:50 am, timinganalyzer <timinganaly...@gmail.com> wrote: > On Jun 29, 11:36 pm, mahesh <mahesh...@gmail.com> wrote: > > > > > On Jun 29, 5:54 pm, timinganalyzer <timinganaly...@gmail.com> wrote: > > > > Hello All, > > > > A new version beta 0.85 is now available. The following changes and > > > additions have occurred. > > > > 1 Quickly add previously used Delays and Constraints from pop-up menu > > > 2 The current state is inverted automatically when adding new pulse if > > > the newstate in the > > > toolbar is the same as the current state in the signal. > > > 3 Keystroke shortcuts for all the signal state buttons in toolbar. > > > Hover over button to see keystrokes. > > > 4 Signals can now use text values for next state. This is useful when > > > labels are needed in digital signals. > > > 5 Moving text and edges by mouse drags now use undo and redo > > > 6 Should work with JRE1.5 or newer > > > > You can download the Free Edition now and read all about the > > > TimingAnalyzer at: > > > >www.timing-diagrams.com > > > Hi, > > I tried to install it on Vista Basic.But facing difficulty in > > installing.I did install the JRE.when I run the timinganalyzer.jar exe > > file It pops me a message saying main class not found exiting !! > > Could u pls help me in fixing the problem ...ur help will be > > appreciated > > > Thanks in advance, > > hesh > > Can you tell what version of Java you are using? I have used beta > 0.85 > with JRE1.5. > > If you bring up a dos window, enter > > java -jar TimingAnalyzer.jar > > This will report the OS, Java version, .... > > Thanks, Dan Hi , I installed java version 1.6.0_06.but unable to run the analyzer...pls do gimme ur inputs... Thanks in advance, heshArticle: 133462
On Jun 30, 1:17 pm, Andy <jonesa...@comcast.net> wrote: > On Jun 30, 11:06 am, Dave <dhsch...@gmail.com> wrote: > > > > > On Jun 30, 11:34 am, Thorsten Kiefer <webmas...@nillakaes.de> wrote: > > > > > 1) Make absolutely sure that you don't have > > > > use ieee.std_logic_arith > > > > or > > > > use ieee.std_logic_[un]signed > > > > in the context clauses. > > > > I don't use them ... > > > > > 2) Just as an experiment (I'm not recommending this for real code) > > > > try > > > > > rnext.addr <= std_logic_vector( > > > > unsigned'(unsigned(rreg.addr) + 1) > > > > ); > > > > rnext.addr <= std_logic_vector( > > > unsigned'(unsigned(rreg.addr) + to_unsigned(18,1) > > > ); > > > works fine. > > > > Regards Thorsten > > > Do you mean this: > > > rnext.addr <= std_logic_vector( > > unsigned'(unsigned(rreg.addr) + to_unsigned(1,18); > > > to_unsigned(18,1) converts the integer 18 to unsigned which is one bit > > wide. I think you meant to convert the integer 1 to an unsigned which > > is 18 bits wide. > > > Also, I don't think you need to cast the result of the addition to > > unsigned, since the result of the addition of two unsigneds is going > > to be an unsigned. There is no ambiguity there. > > > Dave > > > Dave > > You also need not convert the integer to unsigned. Numeric_std "+" is > already defined for unsigned and natural operands (returning > unsigned). > > next.addr <= std_logic_vector(unsigned(rreg.addr) + 1); > > Andy Even though Andy's suggestion is the same as in the original post, he's right. This problem shouldn't be happening if you're using numeric_std, and not using std_logic_arith, std_logic_signed, or std_logic_unsigned. I would re-check your code to make sure those libraries aren't in there. If you're sure, I'd be interested to know what tool it is that's giving you this error. The problem stems from the fact that std_logic_arith has two "+" functions for each possible combination of input argument types: one which has a result of type std_logic_vector, and one which has a result of either signed or unsigned (depending on the input argument types). the compiler doesn't know which one to use. Numeric_std doesn't have this problem, since it only defines the "+" function once for each possible input argument type combination. If, however, you were to use both numeric_std and std_logic_arith, then the ambiguity increases, since there are now 3 options: the two from std_logic_arith, and the one from numeric_std. I notice that your error statement mentions three possibilities... DaveArticle: 133463
On Jun 30, 1:21 pm, mahesh <mahesh...@gmail.com> wrote: > On Jun 30, 5:50 am, timinganalyzer <timinganaly...@gmail.com> wrote: > > > > > On Jun 29, 11:36 pm, mahesh <mahesh...@gmail.com> wrote: > > > > On Jun 29, 5:54 pm, timinganalyzer <timinganaly...@gmail.com> wrote: > > > > > Hello All, > > > > > A new version beta 0.85 is now available. The following changes and > > > > additions have occurred. > > > > > 1 Quickly add previously used Delays and Constraints from pop-up menu > > > > 2 The current state is inverted automatically when adding new pulse if > > > > the newstate in the > > > > toolbar is the same as the current state in the signal. > > > > 3 Keystroke shortcuts for all the signal state buttons in toolbar. > > > > Hover over button to see keystrokes. > > > > 4 Signals can now use text values for next state. This is useful when > > > > labels are needed in digital signals. > > > > 5 Moving text and edges by mouse drags now use undo and redo > > > > 6 Should work with JRE1.5 or newer > > > > > You can download the Free Edition now and read all about the > > > > TimingAnalyzer at: > > > > >www.timing-diagrams.com > > > > Hi, > > > I tried to install it on Vista Basic.But facing difficulty in > > > installing.I did install the JRE.when I run the timinganalyzer.jar exe > > > file It pops me a message saying main class not found exiting !! > > > Could u pls help me in fixing the problem ...ur help will be > > > appreciated > > > > Thanks in advance, > > > hesh > > > Can you tell what version of Java you are using? I have used beta > > 0.85 > > with JRE1.5. > > > If you bring up a dos window, enter > > > java -jar TimingAnalyzer.jar > > > This will report the OS, Java version, .... > > > Thanks, Dan > > Hi , > > I installed java version 1.6.0_06.but unable to run the > analyzer...pls do gimme ur inputs... > > Thanks in advance, > hesh Hello Hesh, Maybe somehow the the zip file you downloaded was corrupted? The file size is 5,170KB. When you extract the zip file, you should see the directory structure shown below c:\Apps\TimingAnalyzer_bxx dir TimingAnalyzer.jar -- The executable program docs dir -- html help files required by program images dir -- images required by program scripts dir -- user scripts themes dir -- look and feel themes pics dir -- saved images of example files examples dir -- timing diagram example files settings dir -- default and user settings If your extracted directory structure and the size the program looks correct, then do the following and let me know the results. Open dos window. cd to TimingAnalyzer_b85 java -jar TimingAnalyzer.jar Copy the output shown in the dos window and send that to me. Let me know how you make out. Thanks, DanArticle: 133464
Hi all, I have two questions: 1.- On the meaning of the syndrome word: Virtex-4 Libraries Guide for HDL Designs says that syndrome(11) = 0 means there is a one bit error and that syndrome(10 downto 0) indicates the error position (pag 94). However, Virtex 4 FPGA Configuration User Guide (pag 75, latest version, april 2008) says that syndrome(11) = 0 means there is more than one bit error and that syndrome(10 downto 0) is meaningless. Does anybody knows which one is correct? 2.- If my design uses LUTRAMs, I am suppose to set the GLUTMASK_B bit to 0 for active readback. By doing this I'll readback all 0's from the bits corresponding to LUTRAMs and I'll avoid any kind of corruption while performing an active readback. This means that whatever I'll be reading back will not be the original bitstream data since the LUTRAM bits are masked. Thus FRAME_ECC_VIRTEX4 will signal an error with every frame readback. Am I right in this assumption? Does anybody have had an experience with this? Any comment would be appreciated, Thanks, Alonzo.Article: 133465
>From pages 75, and 76: http://www.xilinx.com/support/documentation/user_guides/ug071.pdf S[11] = 0, S[10:0] = 0: no error. S[11] = 1, S[10:0] ? 0: single bit (SED) error; S[10:0] denotes location of bit to patch (indirectly). S[11] = 1, S[10:0] = 0: single-bit error; syndrome bit is in error. S[11] = 0, S[10:0] ? 0: double-bit error, not correctable. So, if the MSB (bit 11) equals 1, there is a correctable error (and the bits 10:0 are the indirect location of the bit to flip, or if bits 10:0=0, the error bit itself is in error [ bits 10:0 defined as from 704 = first bit of frame, to 2047 the last bit in the frame]). If the MSB is 0, and the location bits 10:0 are non-zero, then there is a multiple bit hit (uncorrectable). So, for example: 100000000001 -> 640 (bit 0 of syndrome in error) 100000000010 -> 641 100000000100 -> 642 100000001000 -> 643 100000010000 -> 644 100000100000 -> 645 100001000000 -> 646 100010000000 -> 647 100100000000 -> 648 101000000000 -> 649 110000000000 -> 650 100000000000 -> 651 (bit 11 of syndrome in error) all indicate an error in a syndrome bit (there are 11 syndrome bits, which you can correct, too). Austin rha_x@yahoo.com wrote: > Hi all, > I have two questions: > 1.- On the meaning of the syndrome word: Virtex-4 Libraries Guide for > HDL Designs says that syndrome(11) = 0 means there is a one bit error > and that syndrome(10 downto 0) indicates the error position (pag 94). > However, Virtex 4 FPGA Configuration User Guide (pag 75, latest > version, april 2008) says that syndrome(11) = 0 means there is more > than one bit error and that syndrome(10 downto 0) is meaningless. Does > anybody knows which one is correct? > 2.- If my design uses LUTRAMs, I am suppose to set the GLUTMASK_B bit > to 0 for active readback. By doing this I'll readback all 0's from the > bits corresponding to LUTRAMs and I'll avoid any kind of corruption > while performing an active readback. This means that whatever I'll be > reading back will not be the original bitstream data since the LUTRAM > bits are masked. Thus FRAME_ECC_VIRTEX4 will signal an error with > every frame readback. Am I right in this assumption? Does anybody have > had an experience with this? > Any comment would be appreciated, > Thanks, > Alonzo.Article: 133466
If a LUTRAM is used for SRL 16/(32) or for LUTRAM (LUTROM), then the GLUT_Mask bit is set for that LUT, and FRAME_ECC will not include those LUT contents (masked for readback). All other bits in that frame are part of the FRAME_ECC, however. And, yes, the bitstream you read back is not the bitstream you put in. AustinArticle: 133467
Hi, I ran into a problem which is related to TIEOFF_X0Y31. I opened FPGA_Editor and see such sites around Xilinx Virtex5. i am wondering what's this site for? I googled and could not get any infos. Thank you, AdamArticle: 133468
And, I did note inconsistency with the software manual (and reported it to tech pubs). Since I am a "hardware guru" and I know that we write the manuals (first), I have to think the software guide is incorrect. Perhaps that isn't fair, I also know that a '1' in bit 11 indicates an error when the syndrome is non-zero with the location of the bit in error offset by 704. AustinArticle: 133469
On Jun 30, 2:25 pm, austin <aus...@xilinx.com> wrote: > If a LUTRAM is used for SRL 16/(32) or for LUTRAM (LUTROM), then the > GLUT_Mask bit is set for that LUT, and FRAME_ECC will not include those > LUT contents (masked for readback). All other bits in that frame are > part of the FRAME_ECC, however. > > And, yes, the bitstream you read back is not the bitstream you put in. > > Austin Hi Austin, Interesting. I tough FRAME_ECC takes in consideration the ECC bits created during the bitstream generation. If the LUTRAM bits are not taken into account, wouldn't the ECC calculation by FRAME_ECC fail (signal an error). Anyways, I am testing the idea just now. Let's see what happens. I'll post my results. Thank you for your answer. Alonzo.Article: 133470
Hi Raj, This is correct. There is well know result that relates the number of bits you use for quantization with the level of noise added to your signal (about 6dB per bit). Note that quantizing your filter coefficients will result in a completely different filter. There are implementations that are less sensitive to coefficient quantizations than others. Particularly, FIRs are better than IIRs in that sense. Check a DSP book on filters theory. You will find your answers there. Alonzo. . I think it is about 6dB per bit. But don't quote me on that. Check any DSP book. On Jun 30, 12:30 am, rajesho...@yahoo.co.in wrote: > Hai, > > Can i implement FIR filter in FPGA using fixed point number.. > > I referred various FIR filter implementation in FPGA all have used > integer coefficients for their implementation and i am interested to > know the difficulties of using fixed point numbers.I knew fixed point > numbers utilize more hardware. > > I understand fixed point number is scaled and rounded to get integer > values from various tools(FDAtool) hence some information will > lost..Ultimately this will alter the filter > response(cutoff,passband,stopband) > > am i correct? > > if it is possible to implement in fixed point can anyone compare the > accuracy of integer FIR implementation and fixed point implementation? > > pls clarify .. > > regards, > rajArticle: 133471
Adam, There are tie-off pins for the TEMAC, the 440PPC, and other blocks. Which tie-off is this associated with? AustinArticle: 133472
I would like someone to comment my BFSK design as I am a student and not much experienced.Article: 133473
when i try to translate it, it tells me this ERROR:NgdBuild:753 - "top.ucf" Line 4: Could not find instance(s) 'DCM_SP_INST' in the design. To suppress this error specify the correct instance name or remove the constraint. my top.vhd, make use of a component from clockmanager.vhd ...inside clockmanager.vhd has DCM_SP_INST : DCM_SP generic map( CLK_FEEDBACK => "1X", CLKDV_DIVIDE => 2.0, CLKFX_DIVIDE => 1, CLKFX_MULTIPLY => 2, CLKIN_DIVIDE_BY_2 => FALSE, CLKIN_PERIOD => 20.000, CLKOUT_PHASE_SHIFT => "NONE", DESKEW_ADJUST => "SYSTEM_SYNCHRONOUS", DFS_FREQUENCY_MODE => "LOW", DLL_FREQUENCY_MODE => "LOW", DUTY_CYCLE_CORRECTION => TRUE, FACTORY_JF => x"C080", PHASE_SHIFT => 0, STARTUP_WAIT => FALSE) under its architecture what's wrong huh?Article: 133474
kvoskaki@nps.edu wrote: > I would like someone to comment my BFSK design as I am a student and > not much experienced. We can try it. Please upload it to some web host (blog, free web hosting etc.), if it is long and post the link, or post the code (assuming you are using Verilog or VHDL) in this newsgroup, if it is not too long. -- Frank Buss, fb@frank-buss.de http://www.frank-buss.de, http://www.it4-systems.de
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